From 1422fb5d1b89247fa6f99cb4e188fd1dfd47bc4a Mon Sep 17 00:00:00 2001 From: sanaAyrml Date: Thu, 16 Jan 2025 07:02:40 -0500 Subject: [PATCH 1/7] Update scripts --- scripts/eval_scripts/CIFAR10/baseline.slrm | 47 --------------- scripts/eval_scripts/CIFAR10/icgan.slrm | 47 --------------- scripts/eval_scripts/CIFAR10/stablediff.slrm | 47 --------------- scripts/eval_scripts/CIFAR100/baseline.slrm | 48 ---------------- scripts/eval_scripts/CIFAR100/icgan.slrm | 48 ---------------- scripts/eval_scripts/CIFAR100/stablediff.slrm | 48 ---------------- .../eval_scripts/INaturalist/baseline.slrm | 48 ---------------- scripts/eval_scripts/INaturalist/icgan.slrm | 48 ---------------- .../eval_scripts/INaturalist/stablediff.slrm | 48 ---------------- scripts/eval_scripts/food101/baseline.slrm | 47 --------------- scripts/eval_scripts/food101/icgan.slrm | 47 --------------- scripts/eval_scripts/food101/stablediff.slrm | 47 --------------- scripts/eval_scripts/imagenet/baseline.slrm | 45 --------------- scripts/eval_scripts/imagenet/icgan.slrm | 45 --------------- scripts/eval_scripts/imagenet/stablediff.slrm | 45 --------------- scripts/eval_scripts/places365/baseline.slrm | 48 ---------------- scripts/eval_scripts/places365/icgan.slrm | 48 ---------------- .../eval_scripts/places365/stablediff.slrm | 48 ---------------- scripts/generation_scripts/gen_img_icgan.slrm | 4 +- .../gen_img_stablediff.slrm | 4 +- scripts/solo_learn/eval_solo_learn.slrm | 7 +-- scripts/solo_learn/train_solo_learn.slrm | 5 +- .../solo_learn/train_synth_solo_learn.slrm | 28 --------- .../train_simsiam_multinode.slrm | 57 ------------------- .../train_simsiam_singlenode.slrm | 41 ------------- 25 files changed, 9 insertions(+), 986 deletions(-) delete mode 100644 scripts/eval_scripts/CIFAR10/baseline.slrm delete mode 100644 scripts/eval_scripts/CIFAR10/icgan.slrm delete mode 100644 scripts/eval_scripts/CIFAR10/stablediff.slrm delete mode 100644 scripts/eval_scripts/CIFAR100/baseline.slrm delete mode 100644 scripts/eval_scripts/CIFAR100/icgan.slrm delete mode 100644 scripts/eval_scripts/CIFAR100/stablediff.slrm delete mode 100644 scripts/eval_scripts/INaturalist/baseline.slrm delete mode 100644 scripts/eval_scripts/INaturalist/icgan.slrm delete mode 100644 scripts/eval_scripts/INaturalist/stablediff.slrm delete mode 100644 scripts/eval_scripts/food101/baseline.slrm delete mode 100644 scripts/eval_scripts/food101/icgan.slrm delete mode 100644 scripts/eval_scripts/food101/stablediff.slrm delete mode 100644 scripts/eval_scripts/imagenet/baseline.slrm delete mode 100644 scripts/eval_scripts/imagenet/icgan.slrm delete mode 100644 scripts/eval_scripts/imagenet/stablediff.slrm delete mode 100644 scripts/eval_scripts/places365/baseline.slrm delete mode 100644 scripts/eval_scripts/places365/icgan.slrm delete mode 100644 scripts/eval_scripts/places365/stablediff.slrm delete mode 100644 scripts/solo_learn/train_synth_solo_learn.slrm delete mode 100644 scripts/train_scrpits/train_simsiam_multinode.slrm delete mode 100644 scripts/train_scrpits/train_simsiam_singlenode.slrm diff --git a/scripts/eval_scripts/CIFAR10/baseline.slrm b/scripts/eval_scripts/CIFAR10/baseline.slrm deleted file mode 100644 index 3ecb7a4..0000000 --- a/scripts/eval_scripts/CIFAR10/baseline.slrm +++ /dev/null @@ -1,47 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="cifar" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=slurm-cifar10_baseline_160_%j.out -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_baseline_seed43_bs128_rforig_2024-03-05-12-27/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="cifar10" \ - --num_classes=10 diff --git a/scripts/eval_scripts/CIFAR10/icgan.slrm b/scripts/eval_scripts/CIFAR10/icgan.slrm deleted file mode 100644 index f4bf503..0000000 --- a/scripts/eval_scripts/CIFAR10/icgan.slrm +++ /dev/null @@ -1,47 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="cifar" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=slurm-cifar10_baseline_160_%j.out -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_icgan_seed43_bs128_rforig_2024-03-05-12-52/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="cifar10" \ - --num_classes=10 diff --git a/scripts/eval_scripts/CIFAR10/stablediff.slrm b/scripts/eval_scripts/CIFAR10/stablediff.slrm deleted file mode 100644 index 64361fb..0000000 --- a/scripts/eval_scripts/CIFAR10/stablediff.slrm +++ /dev/null @@ -1,47 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="cifar" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=slurm-cifar10_baseline_160_%j.out -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_stablediff_p0p5_seed43_2024-03-05-13-39/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="cifar10" \ - --num_classes=10 diff --git a/scripts/eval_scripts/CIFAR100/baseline.slrm b/scripts/eval_scripts/CIFAR100/baseline.slrm deleted file mode 100644 index a68be76..0000000 --- a/scripts/eval_scripts/CIFAR100/baseline.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="cifar" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=cifar100_baseline_160_%j.out -#SBATCH --error=cifar100_baseline_160_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_baseline_seed43_bs128_rforig_2024-03-05-12-27/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="cifar100" \ - --num_classes=100 \ No newline at end of file diff --git a/scripts/eval_scripts/CIFAR100/icgan.slrm b/scripts/eval_scripts/CIFAR100/icgan.slrm deleted file mode 100644 index 98a2125..0000000 --- a/scripts/eval_scripts/CIFAR100/icgan.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="cifar" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=cifar100_baseline_160_%j.out -#SBATCH --error=cifar100_baseline_160_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_icgan_seed43_bs128_rforig_2024-03-05-12-52/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="cifar100" \ - --num_classes=100 \ No newline at end of file diff --git a/scripts/eval_scripts/CIFAR100/stablediff.slrm b/scripts/eval_scripts/CIFAR100/stablediff.slrm deleted file mode 100644 index 9f6d928..0000000 --- a/scripts/eval_scripts/CIFAR100/stablediff.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="cifar" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=cifar100_baseline_160_%j.out -#SBATCH --error=cifar100_baseline_160_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_stablediff_p0p5_seed43_2024-03-05-13-39/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="cifar100" \ - --num_classes=100 \ No newline at end of file diff --git a/scripts/eval_scripts/INaturalist/baseline.slrm b/scripts/eval_scripts/INaturalist/baseline.slrm deleted file mode 100644 index e68bef7..0000000 --- a/scripts/eval_scripts/INaturalist/baseline.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="inaturalist" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=inaturalist_baseline_%j.out -#SBATCH --error=inaturalist_baseline_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/datasets/inat_comp/2018/" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_baseline_seed43_bs128_rforig_2024-03-05-12-27/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="INaturalist" \ - --num_classes=8142 \ No newline at end of file diff --git a/scripts/eval_scripts/INaturalist/icgan.slrm b/scripts/eval_scripts/INaturalist/icgan.slrm deleted file mode 100644 index 2341e6f..0000000 --- a/scripts/eval_scripts/INaturalist/icgan.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="inaturalist" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=inaturalist_baseline_%j.out -#SBATCH --error=inaturalist_baseline_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/datasets/inat_comp/2018/" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_icgan_seed43_bs128_rforig_2024-03-05-12-52/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="INaturalist" \ - --num_classes=8142 \ No newline at end of file diff --git a/scripts/eval_scripts/INaturalist/stablediff.slrm b/scripts/eval_scripts/INaturalist/stablediff.slrm deleted file mode 100644 index 29f1159..0000000 --- a/scripts/eval_scripts/INaturalist/stablediff.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="inaturalist" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=inaturalist_baseline_%j.out -#SBATCH --error=inaturalist_baseline_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/datasets/inat_comp/2018/" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_stablediff_p0p5_seed43_2024-03-05-13-39/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="INaturalist" \ - --num_classes=8142 \ No newline at end of file diff --git a/scripts/eval_scripts/food101/baseline.slrm b/scripts/eval_scripts/food101/baseline.slrm deleted file mode 100644 index f9f5fdf..0000000 --- a/scripts/eval_scripts/food101/baseline.slrm +++ /dev/null @@ -1,47 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="food101" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=slurm-food101_baseline_160_%j.out -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_baseline_seed43_bs128_rforig_2024-03-05-12-27/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="food101" \ - --num_classes=101 \ No newline at end of file diff --git a/scripts/eval_scripts/food101/icgan.slrm b/scripts/eval_scripts/food101/icgan.slrm deleted file mode 100644 index c31f3a5..0000000 --- a/scripts/eval_scripts/food101/icgan.slrm +++ /dev/null @@ -1,47 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="food101" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=slurm-food101_baseline_160_%j.out -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="P/projects/imagenet_synthetic/model_checkpoints/simsiam_icgan_seed43_bs128_rforig_2024-03-05-12-52/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="food101" \ - --num_classes=101 \ No newline at end of file diff --git a/scripts/eval_scripts/food101/stablediff.slrm b/scripts/eval_scripts/food101/stablediff.slrm deleted file mode 100644 index a30522b..0000000 --- a/scripts/eval_scripts/food101/stablediff.slrm +++ /dev/null @@ -1,47 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="food101" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=slurm-food101_baseline_160_%j.out -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_stablediff_p0p5_seed43_2024-03-05-13-39/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="food101" \ - --num_classes=101 \ No newline at end of file diff --git a/scripts/eval_scripts/imagenet/baseline.slrm b/scripts/eval_scripts/imagenet/baseline.slrm deleted file mode 100644 index 11417ec..0000000 --- a/scripts/eval_scripts/imagenet/baseline.slrm +++ /dev/null @@ -1,45 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="imagenet_eval" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=imagenet_baseline_%j.out -#SBATCH --error=imagenet_baseline_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/scratch/ssd004/datasets/imagenet256" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars --batch-size=2048 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_baseline_seed43_bs128_rforig_2024-03-05-12-27/checkpoint_0160.pth.tar" \ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" diff --git a/scripts/eval_scripts/imagenet/icgan.slrm b/scripts/eval_scripts/imagenet/icgan.slrm deleted file mode 100644 index e68050d..0000000 --- a/scripts/eval_scripts/imagenet/icgan.slrm +++ /dev/null @@ -1,45 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="imagenet_eval" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=imagenet_baseline_%j.out -#SBATCH --error=imagenet_baseline_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/scratch/ssd004/datasets/imagenet256" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars --batch-size=2048 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_icgan_seed43_bs128_rforig_2024-03-05-12-52/checkpoint_0160.pth.tar" \ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" diff --git a/scripts/eval_scripts/imagenet/stablediff.slrm b/scripts/eval_scripts/imagenet/stablediff.slrm deleted file mode 100644 index 37c85c5..0000000 --- a/scripts/eval_scripts/imagenet/stablediff.slrm +++ /dev/null @@ -1,45 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="imagenet_eval" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=imagenet_baseline_%j.out -#SBATCH --error=imagenet_baseline_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/scratch/ssd004/datasets/imagenet256" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars --batch-size=2048 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_stablediff_p0p5_seed43_2024-03-05-13-39/checkpoint_0160.pth.tar" \ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" diff --git a/scripts/eval_scripts/places365/baseline.slrm b/scripts/eval_scripts/places365/baseline.slrm deleted file mode 100644 index a619037..0000000 --- a/scripts/eval_scripts/places365/baseline.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="places365" -#SBATCH --partition=rtx6000 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=places365_baseline_160_%j.out -#SBATCH --error=places365_baseline_160_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets/places365" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_baseline_seed43_bs128_rforig_2024-03-05-12-27/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="places365" \ - --num_classes=434 \ No newline at end of file diff --git a/scripts/eval_scripts/places365/icgan.slrm b/scripts/eval_scripts/places365/icgan.slrm deleted file mode 100644 index 84a9317..0000000 --- a/scripts/eval_scripts/places365/icgan.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="places365" -#SBATCH --partition=rtx6000 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=places365_baseline_160_%j.out -#SBATCH --error=places365_baseline_160_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets/places365" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_icgan_seed43_bs128_rforig_2024-03-05-12-52/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="places365" \ - --num_classes=434 \ No newline at end of file diff --git a/scripts/eval_scripts/places365/stablediff.slrm b/scripts/eval_scripts/places365/stablediff.slrm deleted file mode 100644 index 8985fae..0000000 --- a/scripts/eval_scripts/places365/stablediff.slrm +++ /dev/null @@ -1,48 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="places365" -#SBATCH --partition=rtx6000 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=places365_baseline_160_%j.out -#SBATCH --error=places365_baseline_160_%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -echo $MASTER_ADDR -echo $MASTER_PORT - -export PYTHONPATH="." -nvidia-smi - -python simsiam/linear_eval.py \ - --data="/projects/imagenet_synthetic/fereshteh_datasets/places365" \ - --arch="resnet50" \ - --multiprocessing-distributed \ - --lars \ - --batch-size=4096 \ - --epochs=100 \ - -j=16 \ - --world-size 1 \ - --rank 0 \ - --pretrained="/projects/imagenet_synthetic/model_checkpoints/simsiam_stablediff_p0p5_seed43_2024-03-05-13-39/checkpoint_0160.pth.tar"\ - --dist-url "tcp://$MASTER_ADDR:$MASTER_PORT" \ - --dataset_name="places365" \ - --num_classes=434 \ No newline at end of file diff --git a/scripts/generation_scripts/gen_img_icgan.slrm b/scripts/generation_scripts/gen_img_icgan.slrm index 6741e60..2e9e29e 100644 --- a/scripts/generation_scripts/gen_img_icgan.slrm +++ b/scripts/generation_scripts/gen_img_icgan.slrm @@ -14,7 +14,7 @@ PY_ARGS=${@:1} # activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate +source YOUR_VENV_PATH/bin/activate export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend export CUDA_LAUNCH_BLOCKING=1 @@ -26,7 +26,7 @@ export PYTHONPATH="." nvidia-smi srun python data_generation/img2img_icgan.py \ ---outdir /projects/imagenet_synthetic/synthetic_icgan \ +--outdir SAVE_DIR \ --num_shards=7 \ --shard_index=2 \ --image_version=1 \ diff --git a/scripts/generation_scripts/gen_img_stablediff.slrm b/scripts/generation_scripts/gen_img_stablediff.slrm index 6113e05..87852af 100644 --- a/scripts/generation_scripts/gen_img_stablediff.slrm +++ b/scripts/generation_scripts/gen_img_stablediff.slrm @@ -14,7 +14,7 @@ PY_ARGS=${@:1} # activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate +source YOUR_VENV_PATH/bin/activate export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend export CUDA_LAUNCH_BLOCKING=1 @@ -26,7 +26,7 @@ export PYTHONPATH="." nvidia-smi srun python data_generation/img2img_stable_diff.py \ ---outdir /projects/imagenet_synthetic/arashaf_stablediff_batched \ +--outdir SAVE_DIR \ --num_shards=7 \ --shard_index=2 \ --image_version=1 \ diff --git a/scripts/solo_learn/eval_solo_learn.slrm b/scripts/solo_learn/eval_solo_learn.slrm index 51a5731..0a3666b 100644 --- a/scripts/solo_learn/eval_solo_learn.slrm +++ b/scripts/solo_learn/eval_solo_learn.slrm @@ -1,8 +1,7 @@ #!/bin/bash -#SBATCH --job-name="eval_simsiam_single" -#SBATCH --partition=a40 -#SBATCH --qos=a40_arashaf +#SBATCH --job-name="eval_simclr_single" +#SBATCH --qos=m #SBATCH --nodes=1 #SBATCH --gres=gpu:a40:4 #SBATCH --ntasks-per-node=4 @@ -15,7 +14,7 @@ #SBATCH --time=12:00:00 # load virtual environment -source /ssd003/projects/aieng/envs/genssl3/bin/activate +source YOUR_VENV_PATH/bin/activate export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend diff --git a/scripts/solo_learn/train_solo_learn.slrm b/scripts/solo_learn/train_solo_learn.slrm index fbe9102..9ac8d76 100644 --- a/scripts/solo_learn/train_solo_learn.slrm +++ b/scripts/solo_learn/train_solo_learn.slrm @@ -1,8 +1,7 @@ #!/bin/bash #SBATCH --job-name="simclr_single_train" -#SBATCH --partition=a40 -#SBATCH --qos=a40_arashaf +#SBATCH --qos=m #SBATCH --nodes=1 #SBATCH --gres=gpu:a40:4 #SBATCH --ntasks-per-node=4 @@ -15,7 +14,7 @@ #SBATCH --time=72:00:00 # load virtual environment -source /ssd003/projects/aieng/envs/genssl3/bin/activate +source YOUR_VENV_PATH/bin/activate export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend diff --git a/scripts/solo_learn/train_synth_solo_learn.slrm b/scripts/solo_learn/train_synth_solo_learn.slrm deleted file mode 100644 index 4f11386..0000000 --- a/scripts/solo_learn/train_synth_solo_learn.slrm +++ /dev/null @@ -1,28 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="simclr_single_train" -#SBATCH --partition=a40 -#SBATCH --qos=a40_arashaf -#SBATCH --nodes=1 -#SBATCH --gres=gpu:a40:4 -#SBATCH --ntasks-per-node=4 -#SBATCH --cpus-per-task=8 -#SBATCH --mem=0 -#SBATCH --output=singlenode-%j.out -#SBATCH --error=singlenode-%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=72:00:00 - -# load virtual environment -source /ssd003/projects/aieng/envs/genssl3/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend - -export PYTHONPATH="." -nvidia-smi - -torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_pretrain.py \ - --config-path scripts/pretrain/imagenet/ \ - --config-name simclr_synthetic.yaml \ No newline at end of file diff --git a/scripts/train_scrpits/train_simsiam_multinode.slrm b/scripts/train_scrpits/train_simsiam_multinode.slrm deleted file mode 100644 index 0d4d55c..0000000 --- a/scripts/train_scrpits/train_simsiam_multinode.slrm +++ /dev/null @@ -1,57 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="simsiam_multi_train" -#SBATCH --partition=a40 -#SBATCH --account=deadline -#SBATCH --qos=deadline -#SBATCH --nodes=2 -#SBATCH --gres=gpu:a40:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=01:00:00 -#SBATCH --cpus-per-task=4 -#SBATCH --mem-per-cpu=8G -#SBATCH --output=slurm-%j.out -#SBATCH --error=slurm-%j.err -# load virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 - - -export MASTER_ADDR="$(hostname --fqdn)" -export MASTER_PORT="$(python -c 'import socket; s=socket.socket(); s.bind(("", 0)); print(s.getsockname()[1])')" -export RDVZ_ID=$RANDOM -echo "RDZV Endpoint $MASTER_ADDR:$MASTER_PORT" - -export PYTHONPATH="." -nvidia-smi - -srun -p $SLURM_JOB_PARTITION \ - -c $SLURM_CPUS_ON_NODE \ - -N $SLURM_JOB_NUM_NODES \ - --mem=0 \ - --gres=gpu:$SLURM_JOB_PARTITION:$SLURM_GPUS_ON_NODE \ - bash -c 'torchrun \ - --nproc-per-node=$SLURM_GPUS_ON_NODE \ - --nnodes=$SLURM_JOB_NUM_NODES \ - --rdzv-endpoint $MASTER_ADDR:$MASTER_PORT \ - --rdzv-id $RDVZ_ID \ - --rdzv-backend c10d \ - simsiam/train_simsiam.py.py \ - -a resnet50 \ - --fix-pred-lr \ - --distributed_mode \ - --batch-size=128 \ - --epochs=200 \ - --experiment="simsiam_icgan_seed43_bs128_rforig" \ - --resume_from_checkpoint="/projects/imagenet_synthetic/model_checkpoints/_original_simsiam/checkpoint_0099.pth.tar" \ - --seed=43 \ - --use_synthetic_data \ - --synthetic_data_dir="/projects/imagenet_synthetic/synthetic_icgan" \ - --synthetic_index_min=0 \ - --synthetic_index_max=4 \ - --generative_augmentation_prob=0.5' \ No newline at end of file diff --git a/scripts/train_scrpits/train_simsiam_singlenode.slrm b/scripts/train_scrpits/train_simsiam_singlenode.slrm deleted file mode 100644 index 4be266e..0000000 --- a/scripts/train_scrpits/train_simsiam_singlenode.slrm +++ /dev/null @@ -1,41 +0,0 @@ -#!/bin/bash - -#SBATCH --job-name="simsiam_single_train" -#SBATCH --partition=a40 -#SBATCH --qos=deadline -#SBATCH --account=deadline -#SBATCH --nodes=1 -#SBATCH --gres=gpu:a40:4 -#SBATCH --ntasks-per-node=1 -#SBATCH --cpus-per-task=32 -#SBATCH --mem=0 -#SBATCH --output=singlenode-%j.out -#SBATCH --error=singlenode-%j.err -#SBATCH --open-mode=append -#SBATCH --wait-all-nodes=1 -#SBATCH --time=12:00:00 - -# activate virtual environment -source /ssd003/projects/aieng/envs/genssl2/bin/activate - -export NCCL_IB_DISABLE=1 # Our cluster does not have InfiniBand. We need to disable usage using this flag. -export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 # set to 1 for NCCL backend -# export CUDA_LAUNCH_BLOCKING=1 - -export PYTHONPATH="." -nvidia-smi - -torchrun --nproc-per-node=4 --nnodes=1 simsiam/train_simsiam.py \ - -a resnet50 \ - --fix-pred-lr \ - --distributed_mode \ - --batch-size=128 \ - --epochs=100 \ - --experiment="simsiam_stablediff_p0p5_seed43" \ - --resume_from_checkpoint="" \ - --seed=43 \ - --use_synthetic_data \ - --synthetic_data_dir="/projects/imagenet_synthetic/arashaf_stablediff_batched" \ - --synthetic_index_min=0 \ - --synthetic_index_max=9 \ - --generative_augmentation_prob=0.5 \ No newline at end of file From ce7408e0565f5dde5d44d6209d5ff78bfeeab341 Mon Sep 17 00:00:00 2001 From: sanaAyrml Date: Thu, 16 Jan 2025 07:03:40 -0500 Subject: [PATCH 2/7] Delete simsiam --- simsiam/LARC.py | 107 ----- simsiam/LICENSE | 399 ------------------ simsiam/README.md | 96 ----- simsiam/__init__.py | 0 simsiam/builder.py | 68 --- simsiam/distributed.py | 135 ------ simsiam/inatural_dataset.py | 77 ---- simsiam/linear_eval.py | 807 ------------------------------------ simsiam/loader.py | 124 ------ simsiam/temp.py | 0 simsiam/train_simsiam.py | 438 ------------------- 11 files changed, 2251 deletions(-) delete mode 100644 simsiam/LARC.py delete mode 100644 simsiam/LICENSE delete mode 100644 simsiam/README.md delete mode 100644 simsiam/__init__.py delete mode 100644 simsiam/builder.py delete mode 100644 simsiam/distributed.py delete mode 100644 simsiam/inatural_dataset.py delete mode 100644 simsiam/linear_eval.py delete mode 100644 simsiam/loader.py delete mode 100644 simsiam/temp.py delete mode 100644 simsiam/train_simsiam.py diff --git a/simsiam/LARC.py b/simsiam/LARC.py deleted file mode 100644 index fe41b13..0000000 --- a/simsiam/LARC.py +++ /dev/null @@ -1,107 +0,0 @@ -import torch -from torch import nn -from torch.nn.parameter import Parameter - - -class LARC(object): - """ - :class:`LARC` is a pytorch implementation of both the scaling and clipping variants of LARC, - in which the ratio between gradient and parameter magnitudes is used to calculate an adaptive - local learning rate for each individual parameter. The algorithm is designed to improve - convergence of large batch training. - - See https://arxiv.org/abs/1708.03888 for calculation of the local learning rate. - In practice it modifies the gradients of parameters as a proxy for modifying the learning rate - of the parameters. This design allows it to be used as a wrapper around any torch.optim Optimizer. - ``` - model = ... - optim = torch.optim.Adam(model.parameters(), lr=...) - optim = LARC(optim) - ``` - It can even be used in conjunction with apex.fp16_utils.FP16_optimizer. - ``` - model = ... - optim = torch.optim.Adam(model.parameters(), lr=...) - optim = LARC(optim) - optim = apex.fp16_utils.FP16_Optimizer(optim) - ``` - Args: - optimizer: Pytorch optimizer to wrap and modify learning rate for. - trust_coefficient: Trust coefficient for calculating the lr. See https://arxiv.org/abs/1708.03888 - clip: Decides between clipping or scaling mode of LARC. If `clip=True` the learning rate is set to `min(optimizer_lr, local_lr)` for each parameter. If `clip=False` the learning rate is set to `local_lr*optimizer_lr`. - eps: epsilon kludge to help with numerical stability while calculating adaptive_lr - """ - - def __init__(self, optimizer, trust_coefficient=0.02, clip=True, eps=1e-8): - self.optim = optimizer - self.trust_coefficient = trust_coefficient - self.eps = eps - self.clip = clip - - def __getstate__(self): - return self.optim.__getstate__() - - def __setstate__(self, state): - self.optim.__setstate__(state) - - @property - def state(self): - return self.optim.state - - def __repr__(self): - return self.optim.__repr__() - - @property - def param_groups(self): - return self.optim.param_groups - - @param_groups.setter - def param_groups(self, value): - self.optim.param_groups = value - - def state_dict(self): - return self.optim.state_dict() - - def load_state_dict(self, state_dict): - self.optim.load_state_dict(state_dict) - - def zero_grad(self): - self.optim.zero_grad() - - def add_param_group(self, param_group): - self.optim.add_param_group(param_group) - - def step(self): - with torch.no_grad(): - weight_decays = [] - for group in self.optim.param_groups: - # absorb weight decay control from optimizer - weight_decay = group["weight_decay"] if "weight_decay" in group else 0 - weight_decays.append(weight_decay) - group["weight_decay"] = 0 - for p in group["params"]: - if p.grad is None: - continue - param_norm = torch.norm(p.data) - grad_norm = torch.norm(p.grad.data) - - if param_norm != 0 and grad_norm != 0: - # calculate adaptive lr + weight decay - adaptive_lr = ( - self.trust_coefficient - * (param_norm) - / (grad_norm + param_norm * weight_decay + self.eps) - ) - - # clip learning rate for LARC - if self.clip: - # calculation of adaptive_lr so that when multiplied by lr it equals `min(adaptive_lr, lr)` - adaptive_lr = min(adaptive_lr / group["lr"], 1) - - p.grad.data += weight_decay * p.data - p.grad.data *= adaptive_lr - - self.optim.step() - # return weight decay control to optimizer - for i, group in enumerate(self.optim.param_groups): - group["weight_decay"] = weight_decays[i] diff --git a/simsiam/LICENSE b/simsiam/LICENSE deleted file mode 100644 index 105a4fb..0000000 --- a/simsiam/LICENSE +++ /dev/null @@ -1,399 +0,0 @@ -Attribution-NonCommercial 4.0 International - -======================================================================= - -Creative Commons Corporation ("Creative Commons") is not a law firm and -does not provide legal services or legal advice. Distribution of -Creative Commons public licenses does not create a lawyer-client or -other relationship. Creative Commons makes its licenses and related -information available on an "as-is" basis. Creative Commons gives no -warranties regarding its licenses, any material licensed under their -terms and conditions, or any related information. Creative Commons -disclaims all liability for damages resulting from their use to the -fullest extent possible. - -Using Creative Commons Public Licenses - -Creative Commons public licenses provide a standard set of terms and -conditions that creators and other rights holders may use to share -original works of authorship and other material subject to copyright -and certain other rights specified in the public license below. The -following considerations are for informational purposes only, are not -exhaustive, and do not form part of our licenses. - - Considerations for licensors: Our public licenses are - intended for use by those authorized to give the public - permission to use material in ways otherwise restricted by - copyright and certain other rights. Our licenses are - irrevocable. Licensors should read and understand the terms - and conditions of the license they choose before applying it. - Licensors should also secure all rights necessary before - applying our licenses so that the public can reuse the - material as expected. Licensors should clearly mark any - material not subject to the license. This includes other CC- - licensed material, or material used under an exception or - limitation to copyright. More considerations for licensors: - wiki.creativecommons.org/Considerations_for_licensors - - Considerations for the public: By using one of our public - licenses, a licensor grants the public permission to use the - licensed material under specified terms and conditions. If - the licensor's permission is not necessary for any reason--for - example, because of any applicable exception or limitation to - copyright--then that use is not regulated by the license. Our - licenses grant only permissions under copyright and certain - other rights that a licensor has authority to grant. Use of - the licensed material may still be restricted for other - reasons, including because others have copyright or other - rights in the material. A licensor may make special requests, - such as asking that all changes be marked or described. - Although not required by our licenses, you are encouraged to - respect those requests where reasonable. More_considerations - for the public: - wiki.creativecommons.org/Considerations_for_licensees - -======================================================================= - -Creative Commons Attribution-NonCommercial 4.0 International Public -License - -By exercising the Licensed Rights (defined below), You accept and agree -to be bound by the terms and conditions of this Creative Commons -Attribution-NonCommercial 4.0 International Public License ("Public -License"). To the extent this Public License may be interpreted as a -contract, You are granted the Licensed Rights in consideration of Your -acceptance of these terms and conditions, and the Licensor grants You -such rights in consideration of benefits the Licensor receives from -making the Licensed Material available under these terms and -conditions. - -Section 1 -- Definitions. - - a. Adapted Material means material subject to Copyright and Similar - Rights that is derived from or based upon the Licensed Material - and in which the Licensed Material is translated, altered, - arranged, transformed, or otherwise modified in a manner requiring - permission under the Copyright and Similar Rights held by the - Licensor. For purposes of this Public License, where the Licensed - Material is a musical work, performance, or sound recording, - Adapted Material is always produced where the Licensed Material is - synched in timed relation with a moving image. - - b. Adapter's License means the license You apply to Your Copyright - and Similar Rights in Your contributions to Adapted Material in - accordance with the terms and conditions of this Public License. - - c. Copyright and Similar Rights means copyright and/or similar rights - closely related to copyright including, without limitation, - performance, broadcast, sound recording, and Sui Generis Database - Rights, without regard to how the rights are labeled or - categorized. For purposes of this Public License, the rights - specified in Section 2(b)(1)-(2) are not Copyright and Similar - Rights. - d. Effective Technological Measures means those measures that, in the - absence of proper authority, may not be circumvented under laws - fulfilling obligations under Article 11 of the WIPO Copyright - Treaty adopted on December 20, 1996, and/or similar international - agreements. - - e. Exceptions and Limitations means fair use, fair dealing, and/or - any other exception or limitation to Copyright and Similar Rights - that applies to Your use of the Licensed Material. - - f. Licensed Material means the artistic or literary work, database, - or other material to which the Licensor applied this Public - License. - - g. Licensed Rights means the rights granted to You subject to the - terms and conditions of this Public License, which are limited to - all Copyright and Similar Rights that apply to Your use of the - Licensed Material and that the Licensor has authority to license. - - h. Licensor means the individual(s) or entity(ies) granting rights - under this Public License. - - i. NonCommercial means not primarily intended for or directed towards - commercial advantage or monetary compensation. For purposes of - this Public License, the exchange of the Licensed Material for - other material subject to Copyright and Similar Rights by digital - file-sharing or similar means is NonCommercial provided there is - no payment of monetary compensation in connection with the - exchange. - - j. Share means to provide material to the public by any means or - process that requires permission under the Licensed Rights, such - as reproduction, public display, public performance, distribution, - dissemination, communication, or importation, and to make material - available to the public including in ways that members of the - public may access the material from a place and at a time - individually chosen by them. - - k. Sui Generis Database Rights means rights other than copyright - resulting from Directive 96/9/EC of the European Parliament and of - the Council of 11 March 1996 on the legal protection of databases, - as amended and/or succeeded, as well as other essentially - equivalent rights anywhere in the world. - - l. You means the individual or entity exercising the Licensed Rights - under this Public License. Your has a corresponding meaning. - -Section 2 -- Scope. - - a. License grant. - - 1. Subject to the terms and conditions of this Public License, - the Licensor hereby grants You a worldwide, royalty-free, - non-sublicensable, non-exclusive, irrevocable license to - exercise the Licensed Rights in the Licensed Material to: - - a. reproduce and Share the Licensed Material, in whole or - in part, for NonCommercial purposes only; and - - b. produce, reproduce, and Share Adapted Material for - NonCommercial purposes only. - - 2. Exceptions and Limitations. For the avoidance of doubt, where - Exceptions and Limitations apply to Your use, this Public - License does not apply, and You do not need to comply with - its terms and conditions. - - 3. Term. The term of this Public License is specified in Section - 6(a). - - 4. Media and formats; technical modifications allowed. The - Licensor authorizes You to exercise the Licensed Rights in - all media and formats whether now known or hereafter created, - and to make technical modifications necessary to do so. The - Licensor waives and/or agrees not to assert any right or - authority to forbid You from making technical modifications - necessary to exercise the Licensed Rights, including - technical modifications necessary to circumvent Effective - Technological Measures. For purposes of this Public License, - simply making modifications authorized by this Section 2(a) - (4) never produces Adapted Material. - - 5. Downstream recipients. - - a. Offer from the Licensor -- Licensed Material. Every - recipient of the Licensed Material automatically - receives an offer from the Licensor to exercise the - Licensed Rights under the terms and conditions of this - Public License. - - b. No downstream restrictions. You may not offer or impose - any additional or different terms or conditions on, or - apply any Effective Technological Measures to, the - Licensed Material if doing so restricts exercise of the - Licensed Rights by any recipient of the Licensed - Material. - - 6. No endorsement. Nothing in this Public License constitutes or - may be construed as permission to assert or imply that You - are, or that Your use of the Licensed Material is, connected - with, or sponsored, endorsed, or granted official status by, - the Licensor or others designated to receive attribution as - provided in Section 3(a)(1)(A)(i). - - b. Other rights. - - 1. Moral rights, such as the right of integrity, are not - licensed under this Public License, nor are publicity, - privacy, and/or other similar personality rights; however, to - the extent possible, the Licensor waives and/or agrees not to - assert any such rights held by the Licensor to the limited - extent necessary to allow You to exercise the Licensed - Rights, but not otherwise. - - 2. Patent and trademark rights are not licensed under this - Public License. - - 3. To the extent possible, the Licensor waives any right to - collect royalties from You for the exercise of the Licensed - Rights, whether directly or through a collecting society - under any voluntary or waivable statutory or compulsory - licensing scheme. In all other cases the Licensor expressly - reserves any right to collect such royalties, including when - the Licensed Material is used other than for NonCommercial - purposes. - -Section 3 -- License Conditions. - -Your exercise of the Licensed Rights is expressly made subject to the -following conditions. - - a. Attribution. - - 1. If You Share the Licensed Material (including in modified - form), You must: - - a. retain the following if it is supplied by the Licensor - with the Licensed Material: - - i. identification of the creator(s) of the Licensed - Material and any others designated to receive - attribution, in any reasonable manner requested by - the Licensor (including by pseudonym if - designated); - - ii. a copyright notice; - - iii. a notice that refers to this Public License; - - iv. a notice that refers to the disclaimer of - warranties; - - v. a URI or hyperlink to the Licensed Material to the - extent reasonably practicable; - - b. indicate if You modified the Licensed Material and - retain an indication of any previous modifications; and - - c. indicate the Licensed Material is licensed under this - Public License, and include the text of, or the URI or - hyperlink to, this Public License. - - 2. You may satisfy the conditions in Section 3(a)(1) in any - reasonable manner based on the medium, means, and context in - which You Share the Licensed Material. For example, it may be - reasonable to satisfy the conditions by providing a URI or - hyperlink to a resource that includes the required - information. - - 3. If requested by the Licensor, You must remove any of the - information required by Section 3(a)(1)(A) to the extent - reasonably practicable. - - 4. If You Share Adapted Material You produce, the Adapter's - License You apply must not prevent recipients of the Adapted - Material from complying with this Public License. - -Section 4 -- Sui Generis Database Rights. - -Where the Licensed Rights include Sui Generis Database Rights that -apply to Your use of the Licensed Material: - - a. for the avoidance of doubt, Section 2(a)(1) grants You the right - to extract, reuse, reproduce, and Share all or a substantial - portion of the contents of the database for NonCommercial purposes - only; - - b. if You include all or a substantial portion of the database - contents in a database in which You have Sui Generis Database - Rights, then the database in which You have Sui Generis Database - Rights (but not its individual contents) is Adapted Material; and - - c. You must comply with the conditions in Section 3(a) if You Share - all or a substantial portion of the contents of the database. - -For the avoidance of doubt, this Section 4 supplements and does not -replace Your obligations under this Public License where the Licensed -Rights include other Copyright and Similar Rights. - -Section 5 -- Disclaimer of Warranties and Limitation of Liability. - - a. UNLESS OTHERWISE SEPARATELY UNDERTAKEN BY THE LICENSOR, TO THE - EXTENT POSSIBLE, THE LICENSOR OFFERS THE LICENSED MATERIAL AS-IS - AND AS-AVAILABLE, AND MAKES NO REPRESENTATIONS OR WARRANTIES OF - ANY KIND CONCERNING THE LICENSED MATERIAL, WHETHER EXPRESS, - IMPLIED, STATUTORY, OR OTHER. THIS INCLUDES, WITHOUT LIMITATION, - WARRANTIES OF TITLE, MERCHANTABILITY, FITNESS FOR A PARTICULAR - PURPOSE, NON-INFRINGEMENT, ABSENCE OF LATENT OR OTHER DEFECTS, - ACCURACY, OR THE PRESENCE OR ABSENCE OF ERRORS, WHETHER OR NOT - KNOWN OR DISCOVERABLE. WHERE DISCLAIMERS OF WARRANTIES ARE NOT - ALLOWED IN FULL OR IN PART, THIS DISCLAIMER MAY NOT APPLY TO YOU. - - b. TO THE EXTENT POSSIBLE, IN NO EVENT WILL THE LICENSOR BE LIABLE - TO YOU ON ANY LEGAL THEORY (INCLUDING, WITHOUT LIMITATION, - NEGLIGENCE) OR OTHERWISE FOR ANY DIRECT, SPECIAL, INDIRECT, - INCIDENTAL, CONSEQUENTIAL, PUNITIVE, EXEMPLARY, OR OTHER LOSSES, - COSTS, EXPENSES, OR DAMAGES ARISING OUT OF THIS PUBLIC LICENSE OR - USE OF THE LICENSED MATERIAL, EVEN IF THE LICENSOR HAS BEEN - ADVISED OF THE POSSIBILITY OF SUCH LOSSES, COSTS, EXPENSES, OR - DAMAGES. WHERE A LIMITATION OF LIABILITY IS NOT ALLOWED IN FULL OR - IN PART, THIS LIMITATION MAY NOT APPLY TO YOU. - - c. The disclaimer of warranties and limitation of liability provided - above shall be interpreted in a manner that, to the extent - possible, most closely approximates an absolute disclaimer and - waiver of all liability. - -Section 6 -- Term and Termination. - - a. This Public License applies for the term of the Copyright and - Similar Rights licensed here. However, if You fail to comply with - this Public License, then Your rights under this Public License - terminate automatically. - - b. Where Your right to use the Licensed Material has terminated under - Section 6(a), it reinstates: - - 1. automatically as of the date the violation is cured, provided - it is cured within 30 days of Your discovery of the - violation; or - - 2. upon express reinstatement by the Licensor. - - For the avoidance of doubt, this Section 6(b) does not affect any - right the Licensor may have to seek remedies for Your violations - of this Public License. - - c. For the avoidance of doubt, the Licensor may also offer the - Licensed Material under separate terms or conditions or stop - distributing the Licensed Material at any time; however, doing so - will not terminate this Public License. - - d. Sections 1, 5, 6, 7, and 8 survive termination of this Public - License. - -Section 7 -- Other Terms and Conditions. - - a. The Licensor shall not be bound by any additional or different - terms or conditions communicated by You unless expressly agreed. - - b. Any arrangements, understandings, or agreements regarding the - Licensed Material not stated herein are separate from and - independent of the terms and conditions of this Public License. - -Section 8 -- Interpretation. - - a. For the avoidance of doubt, this Public License does not, and - shall not be interpreted to, reduce, limit, restrict, or impose - conditions on any use of the Licensed Material that could lawfully - be made without permission under this Public License. - - b. To the extent possible, if any provision of this Public License is - deemed unenforceable, it shall be automatically reformed to the - minimum extent necessary to make it enforceable. If the provision - cannot be reformed, it shall be severed from this Public License - without affecting the enforceability of the remaining terms and - conditions. - - c. No term or condition of this Public License will be waived and no - failure to comply consented to unless expressly agreed to by the - Licensor. - - d. Nothing in this Public License constitutes or may be interpreted - as a limitation upon, or waiver of, any privileges and immunities - that apply to the Licensor or You, including from the legal - processes of any jurisdiction or authority. - -======================================================================= - -Creative Commons is not a party to its public -licenses. Notwithstanding, Creative Commons may elect to apply one of -its public licenses to material it publishes and in those instances -will be considered the “Licensor.” The text of the Creative Commons -public licenses is dedicated to the public domain under the CC0 Public -Domain Dedication. Except for the limited purpose of indicating that -material is shared under a Creative Commons public license or as -otherwise permitted by the Creative Commons policies published at -creativecommons.org/policies, Creative Commons does not authorize the -use of the trademark "Creative Commons" or any other trademark or logo -of Creative Commons without its prior written consent including, -without limitation, in connection with any unauthorized modifications -to any of its public licenses or any other arrangements, -understandings, or agreements concerning use of licensed material. For -the avoidance of doubt, this paragraph does not form part of the -public licenses. - -Creative Commons may be contacted at creativecommons.org. \ No newline at end of file diff --git a/simsiam/README.md b/simsiam/README.md deleted file mode 100644 index 47bab1b..0000000 --- a/simsiam/README.md +++ /dev/null @@ -1,96 +0,0 @@ -# SimSiam: Exploring Simple Siamese Representation Learning - -

- simsiam -

- -This is a PyTorch implementation of the [SimSiam paper](https://arxiv.org/abs/2011.10566): -``` -@Article{chen2020simsiam, - author = {Xinlei Chen and Kaiming He}, - title = {Exploring Simple Siamese Representation Learning}, - journal = {arXiv preprint arXiv:2011.10566}, - year = {2020}, -} -``` - -### Preparation - -Install PyTorch and download the ImageNet dataset following the [official PyTorch ImageNet training code](https://github.com/pytorch/examples/tree/master/imagenet). Similar to [MoCo](https://github.com/facebookresearch/moco), the code release contains minimal modifications for both unsupervised pre-training and linear classification to that code. - -In addition, install [apex](https://github.com/NVIDIA/apex) for the [LARS](https://github.com/NVIDIA/apex/blob/master/apex/parallel/LARC.py) implementation needed for linear classification. - -### Unsupervised Pre-Training - -Only **multi-gpu**, **DistributedDataParallel** training is supported; single-gpu or DataParallel training is not supported. - -To do unsupervised pre-training of a ResNet-50 model on ImageNet in an 8-gpu machine, run: -``` -python main_simsiam.py \ - -a resnet50 \ - --dist-url 'tcp://localhost:10001' --multiprocessing-distributed --world-size 1 --rank 0 \ - --fix-pred-lr \ - [your imagenet-folder with train and val folders] -``` -The script uses all the default hyper-parameters as described in the paper, and uses the default augmentation recipe from [MoCo v2](https://arxiv.org/abs/2003.04297). - -The above command performs pre-training with a non-decaying predictor learning rate for 100 epochs, corresponding to the last row of Table 1 in the paper. - -### Linear Classification - -With a pre-trained model, to train a supervised linear classifier on frozen features/weights in an 8-gpu machine, run: -``` -python main_lincls.py \ - -a resnet50 \ - --dist-url 'tcp://localhost:10001' --multiprocessing-distributed --world-size 1 --rank 0 \ - --pretrained [your checkpoint path]/checkpoint_0099.pth.tar \ - --lars \ - [your imagenet-folder with train and val folders] -``` - -The above command uses LARS optimizer and a default batch size of 4096. - -### Models and Logs - -Our pre-trained ResNet-50 models and logs: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
pre-train
epochs		batch
size		pre-train
ckpt		pre-train
log		linear cls.
ckpt		linear cls.
log		top-1 acc.
100512linklinklinklink68.1
100256linklinklinklink68.3
- -Settings for the above: 8 NVIDIA V100 GPUs, CUDA 10.1/CuDNN 7.6.5, PyTorch 1.7.0. - -### Transferring to Object Detection - -Same as [MoCo](https://github.com/facebookresearch/moco) for object detection transfer, please see [moco/detection](https://github.com/facebookresearch/moco/tree/master/detection). - - -### License - -This project is under the CC-BY-NC 4.0 license. See [LICENSE](LICENSE) for details. \ No newline at end of file diff --git a/simsiam/__init__.py b/simsiam/__init__.py deleted file mode 100644 index e69de29..0000000 diff --git a/simsiam/builder.py b/simsiam/builder.py deleted file mode 100644 index 423af89..0000000 --- a/simsiam/builder.py +++ /dev/null @@ -1,68 +0,0 @@ -# Copyright (c) Facebook, Inc. and its affiliates. -# All rights reserved. - -# This source code is licensed under the license found in the -# LICENSE file in the root directory of this source tree. - -import torch -import torch.nn as nn - - -class SimSiam(nn.Module): - """ - Build a SimSiam model. - """ - - def __init__(self, base_encoder, dim=2048, pred_dim=512): - """ - dim: feature dimension (default: 2048) - pred_dim: hidden dimension of the predictor (default: 512) - """ - super(SimSiam, self).__init__() - - # create the encoder - # num_classes is the output fc dimension, zero-initialize last BNs - self.encoder = base_encoder(num_classes=dim, zero_init_residual=True) - - # build a 3-layer projector - prev_dim = self.encoder.fc.weight.shape[1] - self.encoder.fc = nn.Sequential( - nn.Linear(prev_dim, prev_dim, bias=False), - nn.BatchNorm1d(prev_dim), - nn.ReLU(inplace=True), # first layer - nn.Linear(prev_dim, prev_dim, bias=False), - nn.BatchNorm1d(prev_dim), - nn.ReLU(inplace=True), # second layer - self.encoder.fc, - nn.BatchNorm1d(dim, affine=False), - ) # output layer - self.encoder.fc[ - 6 - ].bias.requires_grad = False # hack: not use bias as it is followed by BN - - # build a 2-layer predictor - self.predictor = nn.Sequential( - nn.Linear(dim, pred_dim, bias=False), - nn.BatchNorm1d(pred_dim), - nn.ReLU(inplace=True), # hidden layer - nn.Linear(pred_dim, dim), - ) # output layer - - def forward(self, x1, x2): - """ - Input: - x1: first views of images - x2: second views of images - Output: - p1, p2, z1, z2: predictors and targets of the network - See Sec. 3 of https://arxiv.org/abs/2011.10566 for detailed notations - """ - - # compute features for one view - z1 = self.encoder(x1) # NxC - z2 = self.encoder(x2) # NxC - - p1 = self.predictor(z1) # NxC - p2 = self.predictor(z2) # NxC - - return p1, p2, z1.detach(), z2.detach() diff --git a/simsiam/distributed.py b/simsiam/distributed.py deleted file mode 100644 index 53a555d..0000000 --- a/simsiam/distributed.py +++ /dev/null @@ -1,135 +0,0 @@ -"""Utilities for distributed training.""" -import os -import subprocess - -import torch -import torch.distributed as dist - - -def init_distributed_mode( - launcher, - backend, -) -> None: - """Launch distributed training based on given launcher and backend. - - Parameters - ---------- - launcher : {'pytorch', 'slurm'} - Specifies if pytorch launch utitlity (`torchrun`) is being - used or if running on a SLURM cluster. - backend : {'nccl', 'gloo', 'mpi'} - Specifies which backend to use when initializing a process group. - """ - if launcher == "pytorch": - launch_pytorch_dist(backend) - elif launcher == "slurm": - launch_slurm_dist(backend) - else: - raise RuntimeError( - f"Invalid launcher type: {launcher}. Use 'pytorch' or 'slurm'.", - ) - - -def launch_pytorch_dist(backend) -> None: - """Initialize a distributed process group with PyTorch. - - NOTE: This method relies on `torchrun` to set 'MASTER_ADDR', - 'MASTER_PORT', 'RANK', 'WORLD_SIZE' and 'LOCAL_RANK' as environment variables - - Parameters - ---------- - backend : {'nccl', 'gloo', 'mpi'} - Specifies which backend to use when initializing a process group. Can be - one of ``"nccl"``, ``"gloo"``, or ``"mpi"``. - """ - local_rank = int(os.environ["LOCAL_RANK"]) - torch.cuda.set_device(local_rank) - dist.init_process_group(backend=backend, init_method="env://") - disable_non_master_print() # only print in master process - dist.barrier() - - -def launch_slurm_dist(backend) -> None: - """Initialize a distributed process group when using SLURM. - - Parameters - ---------- - backend : {'nccl', 'gloo', 'mpi'} - Specifies which backend to use when initializing a process group. Can be - one of ``"nccl"``, ``"gloo"``, or ``"mpi"``. - """ - # set the MASTER_ADDR, MASTER_PORT, RANK and WORLD_SIZE - # as environment variables before initializing the process group - if "MASTER_ADDR" not in os.environ: - node_list = os.environ["SLURM_NODELIST"] - os.environ["MASTER_ADDR"] = subprocess.getoutput( - f"scontrol show hostname {node_list} | head -n1", - ) - if "MASTER_PORT" not in os.environ: - os.environ["MASTER_PORT"] = "29400" - os.environ["RANK"] = os.environ["SLURM_PROCID"] - os.environ["WORLD_SIZE"] = os.environ["SLURM_NTASKS"] - - local_rank = int(os.environ["SLURM_LOCALID"]) - print(f"Initializing distributed training in process {local_rank}") - torch.cuda.set_device(local_rank) - dist.init_process_group(backend=backend, init_method="env://") - disable_non_master_print() # only print on master process - dist.barrier() - - -# the following functions were adapted from: -# https://github.com/pytorch/vision/blob/main/references/classification/utils.py -def disable_non_master_print(): - """Disable printing if not master process. - - Notes - ----- - Printing can be forced by adding a boolean flag, 'force', to the keyword arguments - to the print function call. - """ - import builtins as __builtin__ - - builtin_print = __builtin__.print - - def print(*args, **kwargs): # noqa: A001 - force = kwargs.pop("force", False) - if is_main_process() or force: - builtin_print(*args, **kwargs) - - __builtin__.print = print - - -def is_dist_avail_and_initialized() -> bool: - """Check if the distributed package is available and initialized.""" - return dist.is_available() and dist.is_initialized() - - -def get_world_size() -> int: - """Get the total number of processes a distributed process group. - - It returns 1 if the PyTorch distributed package is unavailable or the - default process group has not been initialized. - """ - if not is_dist_avail_and_initialized(): - return 1 - return dist.get_world_size() - - -def get_rank() -> int: - """Return the global rank of the current process. - - Returns 0 if the PyTorch distribued package is unavailable or the - default process group has not been initialized. - """ - if not is_dist_avail_and_initialized(): - return 0 - return dist.get_rank() - - -def is_main_process() -> bool: - """Check if the current process is the Master proces. - - The master process typically has a rank of 0. - """ - return not is_dist_avail_and_initialized() or get_rank() == 0 diff --git a/simsiam/inatural_dataset.py b/simsiam/inatural_dataset.py deleted file mode 100644 index 65b689f..0000000 --- a/simsiam/inatural_dataset.py +++ /dev/null @@ -1,77 +0,0 @@ -import torch.utils.data as data -from PIL import Image -import os -import json -from torchvision import transforms -import random -import numpy as np - - -def default_loader(path): - return Image.open(path).convert("RGB") - - -def load_taxonomy(ann_data, tax_levels, classes): - # loads the taxonomy data and converts to ints - taxonomy = {} - - if "categories" in ann_data.keys(): - num_classes = len(ann_data["categories"]) - for tt in tax_levels: - tax_data = [aa[tt] for aa in ann_data["categories"]] - _, tax_id = np.unique(tax_data, return_inverse=True) - taxonomy[tt] = dict(zip(range(num_classes), list(tax_id))) - else: - # set up dummy data - for tt in tax_levels: - taxonomy[tt] = dict(zip([0], [0])) - - # create a dictionary of lists containing taxonomic labels - classes_taxonomic = {} - for cc in np.unique(classes): - tax_ids = [0] * len(tax_levels) - for ii, tt in enumerate(tax_levels): - tax_ids[ii] = taxonomy[tt][cc] - classes_taxonomic[cc] = tax_ids - - return taxonomy, classes_taxonomic - - -class INAT(data.Dataset): - def __init__(self, root, ann_file, transform): - # load annotations - print("Loading annotations from: " + os.path.basename(ann_file)) - with open(ann_file) as data_file: - ann_data = json.load(data_file) - - # set up the filenames and annotations - self.imgs = [aa["file_name"] for aa in ann_data["images"]] - self.ids = [aa["id"] for aa in ann_data["images"]] - - # if we dont have class labels set them to '0' - if "annotations" in ann_data.keys(): - self.classes = [aa["category_id"] for aa in ann_data["annotations"]] - else: - self.classes = [0] * len(self.imgs) - - # print out some stats - print("\t" + str(len(self.imgs)) + " images") - print("\t" + str(len(set(self.classes))) + " classes") - - self.root = root - self.loader = default_loader - - # augmentation params - self.transform = transform - - def __getitem__(self, index): - path = self.root + self.imgs[index] - img = self.loader(path) - species_id = self.classes[index] - - img = self.transform(img) - - return img, species_id - - def __len__(self): - return len(self.imgs) diff --git a/simsiam/linear_eval.py b/simsiam/linear_eval.py deleted file mode 100644 index e42c097..0000000 --- a/simsiam/linear_eval.py +++ /dev/null @@ -1,807 +0,0 @@ -#!/usr/bin/env python -# Copyright (c) Facebook, Inc. and its affiliates. -# All rights reserved. - -# This source code is licensed under the license found in the -# LICENSE file in the root directory of this source tree. - -import argparse -import builtins -import math -import os -import random -import shutil -import time -import warnings -from datetime import datetime - -import torch -import torch.backends.cudnn as cudnn -import torch.distributed as dist -import torch.multiprocessing as mp -import torch.nn as nn -import torch.nn.parallel -import torch.optim -import torch.utils.data -import torch.utils.data.distributed -import torchvision.datasets as datasets -import torchvision.models as models -import torchvision.transforms as transforms -from tqdm import tqdm -from data_generation.icgan.data_utils import utils as data_utils - -from inatural_dataset import INAT - - -model_names = sorted( - name - for name in models.__dict__ - if name.islower() and not name.startswith("__") and callable(models.__dict__[name]) -) - -parser = argparse.ArgumentParser(description="PyTorch ImageNet Training") -parser.add_argument( - "--data", - metavar="DIR", - default="/scratch/ssd004/datasets/imagenet256", - help="path to dataset.", -) -parser.add_argument( - "-a", - "--arch", - metavar="ARCH", - default="resnet50", - choices=model_names, - help="model architecture: " + " | ".join(model_names) + " (default: resnet50)", -) -parser.add_argument( - "-j", - "--workers", - default=4, - type=int, - metavar="N", - help="number of data loading workers (default: 32)", -) -parser.add_argument( - "--epochs", default=90, type=int, metavar="N", help="number of total epochs to run" -) -parser.add_argument( - "-b", - "--batch-size", - default=4096, - type=int, - metavar="N", - help="mini-batch size (default: 4096), this is the total " - "batch size of all GPUs on the current node when " - "using Data Parallel or Distributed Data Parallel", -) -parser.add_argument( - "--lr", - "--learning-rate", - default=0.1, - type=float, - metavar="LR", - help="initial (base) learning rate", - dest="lr", -) -parser.add_argument("--momentum", default=0.9, type=float, metavar="M", help="momentum") -parser.add_argument( - "--wd", - "--weight-decay", - default=0.0, - type=float, - metavar="W", - help="weight decay (default: 0.)", - dest="weight_decay", -) -parser.add_argument( - "-p", - "--print-freq", - default=10, - type=int, - metavar="N", - help="print frequency (default: 10)", -) -parser.add_argument( - "--resume", - default="", - type=str, - metavar="PATH", - help="path to latest checkpoint (default: none)", -) -parser.add_argument( - "-e", - "--evaluate", - dest="evaluate", - action="store_true", - help="evaluate model on validation set", -) -parser.add_argument( - "--world-size", - default=-1, - type=int, - help="number of nodes for distributed training", -) -parser.add_argument( - "--rank", default=-1, type=int, help="node rank for distributed training" -) -parser.add_argument( - "--dist-url", - default="tcp://224.66.41.62:23456", - type=str, - help="url used to set up distributed training", -) -parser.add_argument( - "--dist-backend", default="nccl", type=str, help="distributed backend" -) -parser.add_argument( - "--seed", default=None, type=int, help="seed for initializing training. " -) -parser.add_argument("--gpu", default=None, type=int, help="GPU id to use.") -parser.add_argument( - "--multiprocessing-distributed", - action="store_true", - help="Use multi-processing distributed training to launch " - "N processes per node, which has N GPUs. This is the " - "fastest way to use PyTorch for either single node or " - "multi node data parallel training", -) - -# additional configs: -parser.add_argument( - "--pretrained", default="", type=str, help="path to simsiam pretrained checkpoint" -) -parser.add_argument("--lars", action="store_true", help="Use LARS") - -parser.add_argument("--dataset_name", default="imagenet", help="Name of the dataset.") - -parser.add_argument( - "--checkpoint_dir", - default="/projects/imagenet_synthetic/model_checkpoints", - help="Checkpoint root directory.", -) - -parser.add_argument( - "--num_classes", - default=1000, - type=int, - help="Number of classes in the dataset.", -) - -parser.add_argument( - "--ablation_mode", - default="icgan", - type=str, - help="Using icgan or stable diffusion feature extractor for ablation study.", -) - -best_acc1 = 0 - - -def main(): - args = parser.parse_args() - current_time = datetime.now().strftime("%Y-%m-%d-%H-%M") - args.checkpoint_dir = os.path.join(args.checkpoint_dir, f"eval_{current_time}") - os.makedirs(args.checkpoint_dir, exist_ok=True) - - print(args) - - if args.seed is not None: - random.seed(args.seed) - torch.manual_seed(args.seed) - # NOTE: this line can reduce speed considerably - cudnn.deterministic = True - warnings.warn( - "You have chosen to seed training. " - "This will turn on the CUDNN deterministic setting, " - "which can slow down your training considerably! " - "You may see unexpected behavior when restarting " - "from checkpoints." - ) - - if args.gpu is not None: - warnings.warn( - "You have chosen a specific GPU. This will completely " - "disable data parallelism." - ) - - if args.dist_url == "env://" and args.world_size == -1: - args.world_size = int(os.environ["WORLD_SIZE"]) - args.distributed = args.world_size > 1 or args.multiprocessing_distributed - - ngpus_per_node = torch.cuda.device_count() - if args.multiprocessing_distributed: - # Since we have ngpus_per_node processes per node, the total world_size - # needs to be adjusted accordingly - args.world_size = ngpus_per_node * args.world_size - # Use torch.multiprocessing.spawn to launch distributed processes: the - # main_worker process function - mp.spawn( - main_worker, - nprocs=ngpus_per_node, - args=( - ngpus_per_node, - args, - ), - ) - else: - # Simply call main_worker function - main_worker(args.gpu, ngpus_per_node, args) - - -def main_worker(gpu, ngpus_per_node, args): - global best_acc1 - args.gpu = gpu - - # suppress printing if not master - if args.multiprocessing_distributed and args.gpu != 0: - - def print_pass(*args, flush=True): - pass - - builtins.print = print_pass - - if args.gpu is not None: - print("Use GPU: {} for training".format(args.gpu), flush=True) - - if args.distributed: - if args.dist_url == "env://" and args.rank == -1: - args.rank = int(os.environ["RANK"]) - if args.multiprocessing_distributed: - # For multiprocessing distributed training, rank needs to be the - # global rank among all the processes - args.rank = args.rank * ngpus_per_node + gpu - dist.init_process_group( - backend=args.dist_backend, - init_method=args.dist_url, - world_size=args.world_size, - rank=args.rank, - ) - torch.distributed.barrier() - - # create model - print("=> creating model '{}'".format(args.arch), flush=True) - model = models.__dict__[args.arch]() - - model.fc = nn.Linear(2048, args.num_classes) - - # freeze all layers but the last fc - for name, param in model.named_parameters(): - if name not in ["fc.weight", "fc.bias"]: - param.requires_grad = False - # init the fc layer - model.fc.weight.data.normal_(mean=0.0, std=0.01) - model.fc.bias.data.zero_() - - # load from pre-trained, before DistributedDataParallel constructor - if args.pretrained: - if os.path.isfile(args.pretrained): - print("=> loading checkpoint '{}'".format(args.pretrained), flush=True) - checkpoint = torch.load(args.pretrained, map_location="cpu") - - # rename moco pre-trained keys - state_dict = checkpoint["state_dict"] - for k in list(state_dict.keys()): - # retain only encoder up to before the embedding layer - if k.startswith("module.encoder") and not k.startswith( - "module.encoder.fc" - ): - # remove prefix - state_dict[k[len("module.encoder.") :]] = state_dict[k] - # delete renamed or unused k - del state_dict[k] - - args.start_epoch = 0 - msg = model.load_state_dict(state_dict, strict=False) - assert set(msg.missing_keys) == {"fc.weight", "fc.bias"} - - print("=> loaded pre-trained model '{}'".format(args.pretrained)) - else: - print("=> no checkpoint found at '{}'".format(args.pretrained)) - - # infer learning rate before changing batch size - init_lr = args.lr * args.batch_size / 256 - - if args.distributed: - # For multiprocessing distributed, DistributedDataParallel constructor - # should always set the single device scope, otherwise, - # DistributedDataParallel will use all available devices. - if args.gpu is not None: - torch.cuda.set_device(args.gpu) - model.cuda(args.gpu) - # When using a single GPU per process and per - # DistributedDataParallel, we need to divide the batch size - # ourselves based on the total number of GPUs we have - args.batch_size = int(args.batch_size / ngpus_per_node) - args.workers = int((args.workers + ngpus_per_node - 1) / ngpus_per_node) - model = torch.nn.parallel.DistributedDataParallel( - model, device_ids=[args.gpu] - ) - else: - model.cuda() - # DistributedDataParallel will divide and allocate batch_size to all - # available GPUs if device_ids are not set - model = torch.nn.parallel.DistributedDataParallel(model) - elif args.gpu is not None: - torch.cuda.set_device(args.gpu) - model = model.cuda(args.gpu) - else: - # DataParallel will divide and allocate batch_size to all available GPUs - if args.arch.startswith("alexnet") or args.arch.startswith("vgg"): - model.features = torch.nn.DataParallel(model.features) - model.cuda() - else: - model = torch.nn.DataParallel(model).cuda() - - # define loss function (criterion) and optimizer - criterion = nn.CrossEntropyLoss().cuda(args.gpu) - - # optimize only the linear classifier - parameters = list(filter(lambda p: p.requires_grad, model.parameters())) - assert len(parameters) == 2 # fc.weight, fc.bias - - optimizer = torch.optim.SGD( - parameters, init_lr, momentum=args.momentum, weight_decay=args.weight_decay - ) - if args.lars: - print("=> use LARS optimizer.", flush=True) - from LARC import LARC - - optimizer = LARC(optimizer=optimizer, trust_coefficient=0.001, clip=False) - - # optionally resume from a checkpoint - if args.resume: - if os.path.isfile(args.resume): - print("=> loading checkpoint '{}'".format(args.resume), flush=True) - if args.gpu is None: - checkpoint = torch.load(args.resume) - else: - # Map model to be loaded to specified single gpu. - loc = "cuda:{}".format(args.gpu) - checkpoint = torch.load(args.resume, map_location=loc) - args.start_epoch = checkpoint["epoch"] - best_acc1 = checkpoint["best_acc1"] - if args.gpu is not None: - # best_acc1 may be from a checkpoint from a different GPU - best_acc1 = best_acc1.to(args.gpu) - model.load_state_dict(checkpoint["state_dict"]) - optimizer.load_state_dict(checkpoint["optimizer"]) - print( - "=> loaded checkpoint '{}' (epoch {})".format( - args.resume, checkpoint["epoch"] - ), - flush=True, - ) - else: - print("=> no checkpoint found at '{}'".format(args.resume)) - - cudnn.benchmark = True - - # Data loading code - traindir = os.path.join(args.data, "train") - valdir = os.path.join(args.data, "val") - normalize = transforms.Normalize( - mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225] - ) - - if args.dataset_name == "imagenet": - train_dataset = datasets.ImageFolder( - traindir, - transforms.Compose( - [ - transforms.RandomResizedCrop(224), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - normalize, - ] - ), - ) - val_dataset = datasets.ImageFolder( - valdir, - transforms.Compose( - [ - transforms.Resize(256), - transforms.CenterCrop(224), - transforms.ToTensor(), - normalize, - ] - ), - ) - elif args.dataset_name == "food101": - print("=> using food101 dataset.", flush=True) - train_dataset = datasets.Food101( - root=args.data, - split="train", - transform=transforms.Compose( - [ - transforms.RandomResizedCrop(224), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - normalize, - ], - ), - ) - val_dataset = datasets.Food101( - root=args.data, - split="test", - transform=transforms.Compose( - [ - transforms.Resize(256), - transforms.CenterCrop(224), - transforms.ToTensor(), - normalize, - ], - ), - ) - elif args.dataset_name == "cifar10": - train_dataset = datasets.CIFAR10( - root=args.data, - train=True, - download=True, - transform=transforms.Compose( - [ - transforms.RandomResizedCrop(224), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)), - ], - ), - ) - val_dataset = datasets.CIFAR10( - root=args.data, - train=False, - download=True, - transform=transforms.Compose( - [ - transforms.Resize(256), - transforms.CenterCrop(224), - transforms.ToTensor(), - transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)), - ], - ), - ) - elif args.dataset_name == "cifar100": - train_dataset = datasets.CIFAR100( - root=args.data, - train=True, - transform=transforms.Compose( - [ - transforms.RandomResizedCrop(224), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)), - ], - ), - ) - val_dataset = datasets.CIFAR100( - root=args.data, - train=False, - transform=transforms.Compose( - [ - transforms.Resize(256), - transforms.CenterCrop(224), - transforms.ToTensor(), - transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)), - ], - ), - ) - elif args.dataset_name == "places365": - train_dataset = datasets.Places365( - root=args.data, - split="train-standard", - transform=transforms.Compose( - [ - transforms.RandomResizedCrop(224), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - normalize, - ], - ), - ) - val_dataset = datasets.Places365( - root=args.data, - split="val", - transform=transforms.Compose( - [ - transforms.Resize(256), - transforms.CenterCrop(224), - transforms.ToTensor(), - normalize, - ], - ), - ) - elif args.dataset_name == "INaturalist": - train_dataset = INAT( - root=args.data, - ann_file=os.path.join(args.data, "train2018.json"), - transform=transforms.Compose( - [ - transforms.RandomResizedCrop(224), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - normalize, - ], - ), - ) - val_dataset = INAT( - root=args.data, - ann_file=os.path.join(args.data, "val2018.json"), - transform=transforms.Compose( - [ - transforms.Resize(256), - transforms.CenterCrop(224), - transforms.ToTensor(), - normalize, - ], - ), - ) - - if args.distributed: - train_sampler = torch.utils.data.distributed.DistributedSampler(train_dataset) - else: - train_sampler = None - - train_loader = torch.utils.data.DataLoader( - train_dataset, - batch_size=args.batch_size, - shuffle=(train_sampler is None), - num_workers=args.workers, - pin_memory=True, - sampler=train_sampler, - ) - - val_loader = torch.utils.data.DataLoader( - val_dataset, - batch_size=256, - shuffle=False, - num_workers=args.workers, - pin_memory=True, - ) - - if args.evaluate: - validate(val_loader, model, criterion, args) - return - - for epoch in range(args.start_epoch, args.epochs): - if args.distributed: - train_sampler.set_epoch(epoch) - adjust_learning_rate(optimizer, init_lr, epoch, args) - - # train for one epoch - train(train_loader, model, criterion, optimizer, epoch, args) - - # evaluate on validation set - acc1 = validate(val_loader, model, criterion, args) - - # remember best acc@1 and save checkpoint - is_best = acc1 > best_acc1 - best_acc1 = max(acc1, best_acc1) - - if not args.multiprocessing_distributed or ( - args.multiprocessing_distributed and args.rank % ngpus_per_node == 0 - ): - checkpoint_name = "checkpoint_{:04d}.pth.tar".format(epoch + 1) - checkpoint_file = os.path.join(args.checkpoint_dir, checkpoint_name) - save_checkpoint( - { - "epoch": epoch + 1, - "arch": args.arch, - "state_dict": model.state_dict(), - "best_acc1": best_acc1, - "optimizer": optimizer.state_dict(), - }, - is_best, - filename=checkpoint_file, - ) - if epoch == args.start_epoch: - sanity_check(model.state_dict(), args.pretrained) - - -def train(train_loader, model, criterion, optimizer, epoch, args): - batch_time = AverageMeter("Time", ":6.3f") - data_time = AverageMeter("Data", ":6.3f") - losses = AverageMeter("Loss", ":.4e") - top1 = AverageMeter("Acc@1", ":6.2f") - top5 = AverageMeter("Acc@5", ":6.2f") - progress = ProgressMeter( - len(train_loader), - [batch_time, data_time, losses, top1, top5], - prefix="Epoch: [{}]".format(epoch), - ) - - """ - Switch to eval mode: - Under the protocol of linear classification on frozen features/models, - it is not legitimate to change any part of the pre-trained model. - BatchNorm in train mode may revise running mean/std (even if it receives - no gradient), which are part of the model parameters too. - """ - model.eval() - - end = time.time() - i = 0 - for images, target in tqdm(train_loader): - # measure data loading time - data_time.update(time.time() - end) - - if args.gpu is not None: - images = images.cuda(args.gpu, non_blocking=True) - target = target.cuda(args.gpu, non_blocking=True) - - # compute output - output = model(images) - loss = criterion(output, target) - - # measure accuracy and record loss - acc1, acc5 = accuracy(output, target, topk=(1, 5)) - losses.update(loss.item(), images.size(0)) - top1.update(acc1[0], images.size(0)) - top5.update(acc5[0], images.size(0)) - - # compute gradient and do SGD step - optimizer.zero_grad() - loss.backward() - optimizer.step() - - # measure elapsed time - batch_time.update(time.time() - end) - end = time.time() - - if i % args.print_freq == 0: - progress.display(i) - - i += 1 - - -def validate(val_loader, model, criterion, args): - batch_time = AverageMeter("Time", ":6.3f") - losses = AverageMeter("Loss", ":.4e") - top1 = AverageMeter("Acc@1", ":6.2f") - top5 = AverageMeter("Acc@5", ":6.2f") - progress = ProgressMeter( - len(val_loader), [batch_time, losses, top1, top5], prefix="Test: " - ) - - # switch to evaluate mode - model.eval() - - with torch.no_grad(): - end = time.time() - i = 0 - for images, target in tqdm(val_loader): - if args.gpu is not None: - images = images.cuda(args.gpu, non_blocking=True) - target = target.cuda(args.gpu, non_blocking=True) - - # compute output - output = model(images) - loss = criterion(output, target) - - # measure accuracy and record loss - acc1, acc5 = accuracy(output, target, topk=(1, 5)) - losses.update(loss.item(), images.size(0)) - top1.update(acc1[0], images.size(0)) - top5.update(acc5[0], images.size(0)) - - # measure elapsed time - batch_time.update(time.time() - end) - end = time.time() - - if i % args.print_freq == 0: - progress.display(i) - - i += 1 - - # # TODO: this should also be done with the ProgressMeter - print( - "\n * Accuracy@1 {top1.avg:.3f} Accuracy@5 {top5.avg:.3f}".format( - top1=top1, top5=top5 - ) - ) - - return top1.avg - - -def save_checkpoint(state, is_best, filename="checkpoint.pth.tar"): - torch.save(state, filename) - if is_best: - shutil.copyfile(filename, "model_best.pth.tar") - - -def sanity_check(state_dict, pretrained_weights): - """ - Linear classifier should not change any weights other than the linear layer. - This sanity check asserts nothing wrong happens (e.g., BN stats updated). - """ - print("=> loading '{}' for sanity check".format(pretrained_weights)) - checkpoint = torch.load(pretrained_weights, map_location="cpu") - - state_dict_pre = checkpoint["state_dict"] - - for k in list(state_dict.keys()): - # only ignore fc layer - if "fc.weight" in k or "fc.bias" in k: - continue - - # name in pretrained model - k_pre = ( - "module.encoder." + k[len("module.") :] - if k.startswith("module.") - else "module.encoder." + k - ) - - assert ( - state_dict[k].cpu() == state_dict_pre[k_pre] - ).all(), "{} is changed in linear classifier training.".format(k) - - print("=> sanity check passed.") - - -class AverageMeter(object): - """Computes and stores the average and current value""" - - def __init__(self, name, fmt=":f"): - self.name = name - self.fmt = fmt - self.reset() - - def reset(self): - self.val = 0 - self.avg = 0 - self.sum = 0 - self.count = 0 - - def update(self, val, n=1): - self.val = val - self.sum += val * n - self.count += n - self.avg = self.sum / self.count - - def __str__(self): - fmtstr = "{name} {val" + self.fmt + "} ({avg" + self.fmt + "})" - return fmtstr.format(**self.__dict__) - - -class ProgressMeter(object): - def __init__(self, num_batches, meters, prefix=""): - self.batch_fmtstr = self._get_batch_fmtstr(num_batches) - self.meters = meters - self.prefix = prefix - - def display(self, batch): - entries = [self.prefix + self.batch_fmtstr.format(batch)] - entries += [str(meter) for meter in self.meters] - print("\t".join(entries), flush=True) - - def _get_batch_fmtstr(self, num_batches): - num_digits = len(str(num_batches // 1)) - fmt = "{:" + str(num_digits) + "d}" - return "[" + fmt + "/" + fmt.format(num_batches) + "]" - - -def adjust_learning_rate(optimizer, init_lr, epoch, args): - """Decay the learning rate based on schedule""" - cur_lr = init_lr * 0.5 * (1.0 + math.cos(math.pi * epoch / args.epochs)) - for param_group in optimizer.param_groups: - param_group["lr"] = cur_lr - - -def accuracy(output, target, topk=(1,)): - """Computes the accuracy over the k top predictions for the specified values of k""" - with torch.no_grad(): - maxk = max(topk) - batch_size = target.size(0) - - _, pred = output.topk(maxk, 1, True, True) - pred = pred.t() - correct = pred.eq(target.view(1, -1).expand_as(pred)) - - res = [] - for k in topk: - correct_k = correct[:k].reshape(-1).float().sum(0, keepdim=True) - res.append(correct_k.mul_(100.0 / batch_size)) - return res - - -if __name__ == "__main__": - main() diff --git a/simsiam/loader.py b/simsiam/loader.py deleted file mode 100644 index 5b53049..0000000 --- a/simsiam/loader.py +++ /dev/null @@ -1,124 +0,0 @@ -# Copyright (c) Facebook, Inc. and its affiliates. -# All rights reserved. - -# This source code is licensed under the license found in the -# LICENSE file in the root directory of this source tree. - -import os -import random - -import torch -from PIL import Image, ImageFilter -from torchvision import datasets, transforms - - -class GaussianBlur(object): - """Gaussian blur augmentation in SimCLR https://arxiv.org/abs/2002.05709.""" - - def __init__(self, sigma=[0.1, 2.0]): - self.sigma = sigma - - def __call__(self, x): - sigma = random.uniform(self.sigma[0], self.sigma[1]) - x = x.filter(ImageFilter.GaussianBlur(radius=sigma)) - return x - - -_normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]) - -# MoCo v2's aug: similar to SimCLR https://arxiv.org/abs/2002.05709 -_real_augmentations = [ - transforms.RandomResizedCrop(224, scale=(0.2, 1.0)), - transforms.RandomApply( - [transforms.ColorJitter(0.4, 0.4, 0.4, 0.1)], # not strengthened - p=0.8, - ), - transforms.RandomGrayscale(p=0.2), - transforms.RandomApply([GaussianBlur([0.1, 2.0])], p=0.5), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - _normalize, -] - - -class TwoCropsTransform: - """Take two random crops of one image as the query and key.""" - - def __init__(self): - self.base_transform = transforms.Compose(_real_augmentations) - - def __call__(self, x): - q = self.base_transform(x) - k = self.base_transform(x) - return [q, k] - - -class ImageNetSynthetic(datasets.ImageNet): - def __init__( - self, - imagenet_root, - imagenet_synthetic_root, - index_min=0, - index_max=9, - generative_augmentation_prob=None, - load_one_real_image=False, - split="train", - ): - super(ImageNetSynthetic, self).__init__( - root=imagenet_root, - split=split, - ) - self.imagenet_root = imagenet_root - self.imagenet_synthetic_root = imagenet_synthetic_root - self.index_min = index_min - self.index_max = index_max - self.generative_augmentation_prob = generative_augmentation_prob - self.load_one_real_image = load_one_real_image - self.real_transforms = transforms.Compose(_real_augmentations) - # Remove random crop for synthetic image augmentation. - self.synthetic_transforms = transforms.Compose(_real_augmentations[1:]) - self.split = split - - def __getitem__(self, index): - imagenet_filename, label = self.imgs[index] - - def _synthetic_image(filename): - rand_int = random.randint(self.index_min, self.index_max) - filename_and_extension = filename.split("/")[-1] - filename_parent_dir = filename.split("/")[-2] - image_path = os.path.join( - self.imagenet_synthetic_root, - self.split, - filename_parent_dir, - filename_and_extension.split(".")[0] + f"_{rand_int}.JPEG", - ) - return Image.open(image_path).convert("RGB") - - if self.generative_augmentation_prob is not None: - if torch.rand(1) < self.generative_augmentation_prob: - # Generate a synthetic image. - image1 = _synthetic_image(imagenet_filename) - image1 = self.synthetic_transforms(image1) - else: - image1 = self.loader(os.path.join(self.root, imagenet_filename)) - image1 = self.real_transforms(image1) - - if torch.rand(1) < self.generative_augmentation_prob: - # Generate another synthetic image. - image2 = _synthetic_image(imagenet_filename) - image2 = self.synthetic_transforms(image2) - else: - image2 = self.loader(os.path.join(self.root, imagenet_filename)) - image2 = self.real_transforms(image2) - else: - if self.load_one_real_image: - image1 = self.loader(os.path.join(self.root, imagenet_filename)) - image1 = self.real_transforms(image1) - else: - image1 = _synthetic_image(imagenet_filename) - image1 = self.synthetic_transforms(image1) - # image2 is always synthetic. - image2 = _synthetic_image(imagenet_filename) - image2 = self.synthetic_transforms(image2) - - return [image1, image2], label diff --git a/simsiam/temp.py b/simsiam/temp.py deleted file mode 100644 index e69de29..0000000 diff --git a/simsiam/train_simsiam.py b/simsiam/train_simsiam.py deleted file mode 100644 index 4896c2e..0000000 --- a/simsiam/train_simsiam.py +++ /dev/null @@ -1,438 +0,0 @@ -# Copyright (c) Facebook, Inc. and its affiliates. -# All rights reserved. - -# This source code is licensed under the license found in the -# LICENSE file in the root directory of this source tree. - -import argparse -import math -import os -import random -from datetime import datetime -from functools import partial - -import torch -import torch.nn.parallel -import torch.optim -import torch.utils.data -import torch.utils.data.distributed -from torch import distributed as dist -from torch import nn -from torch.backends import cudnn -from torch.nn.parallel import DistributedDataParallel as DDP # noqa: N817 -from torch.utils.data.distributed import DistributedSampler -from torchvision import datasets, models -from tqdm import tqdm - -from simsiam import distributed as dist_utils -from simsiam import builder, loader - - -model_names = sorted( - name - for name in models.__dict__ - if name.islower() and not name.startswith("__") and callable(models.__dict__[name]) -) - -parser = argparse.ArgumentParser(description="PyTorch ImageNet Training") -parser.add_argument( - "--data_dir", - metavar="DIR", - default="/scratch/ssd004/datasets/imagenet256", - help="path to dataset.", -) -parser.add_argument( - "-a", - "--arch", - metavar="ARCH", - default="resnet50", - choices=model_names, - help="model architecture: " + " | ".join(model_names) + " (default: resnet50)", -) -parser.add_argument( - "-j", - "--num_workers", - default=4, - type=int, - metavar="N", - help="number of data loading workers (default: 32)", -) -parser.add_argument( - "--epochs", default=100, type=int, metavar="N", help="number of total epochs to run" -) -parser.add_argument( - "-b", - "--batch-size", - default=256, - type=int, - metavar="N", - help="mini-batch size (default: 512), this is the total " - "batch size of all GPUs on the current node when " - "using Data Parallel or Distributed Data Parallel", -) -parser.add_argument( - "--lr", - "--learning-rate", - default=0.05, - type=float, - metavar="LR", - help="initial (base) learning rate", - dest="lr", -) -parser.add_argument( - "--momentum", default=0.9, type=float, metavar="M", help="momentum of SGD solver" -) -parser.add_argument( - "--wd", - "--weight-decay", - default=1e-4, - type=float, - metavar="W", - help="weight decay (default: 1e-4)", - dest="weight_decay", -) -parser.add_argument( - "--resume_from_checkpoint", - default="", - type=str, - help="Path to latest checkpoint.", -) -parser.add_argument( - "--seed", default=42, type=int, help="seed for initializing training. " -) - -# simsiam specific configs: -parser.add_argument( - "--dim", default=2048, type=int, help="feature dimension (default: 2048)" -) -parser.add_argument( - "--pred-dim", - default=512, - type=int, - help="hidden dimension of the predictor (default: 512)", -) -parser.add_argument( - "--fix-pred-lr", action="store_true", help="Fix learning rate for the predictor" -) - -parser.add_argument( - "--distributed_mode", - action="store_true", - help="Enable distributed training", -) -parser.add_argument("--distributed_launcher", default="slurm") -parser.add_argument("--distributed_backend", default="nccl") -parser.add_argument( - "--checkpoint_dir", - default="/projects/imagenet_synthetic/model_checkpoints", - help="Checkpoint root directory.", -) -parser.add_argument( - "--experiment", - default="", - help="Experiment name.", -) -parser.add_argument( - "--use_synthetic_data", - action=argparse.BooleanOptionalAction, - help="Whether to use real data or synthetic data for training.", -) -parser.add_argument( - "--synthetic_data_dir", - default="/projects/imagenet_synthetic/", - help="Path to the root of synthetic data.", -) -parser.add_argument( - "--synthetic_index_min", - default=0, - type=int, - help="Synthetic data files are named filename_i.JPEG. This index determines the lower bound for i.", -) -parser.add_argument( - "--synthetic_index_max", - default=9, - type=int, - help="Synthetic data files are named filename_i.JPEG. This index determines the upper bound for i.", -) -parser.add_argument( - "--generative_augmentation_prob", - default=None, - type=float, - help="The probability of applying a generative model augmentation to a view. Applies to the views separately.", -) -parser.add_argument( - "-p", - "--print-freq", - default=10, - type=int, - metavar="N", - help="print frequency (default: 10)", -) - - -def worker_init_fn(worker_id: int, num_workers: int, rank: int, seed: int) -> None: - """Initialize worker processes with a random seed. - - Parameters - ---------- - worker_id : int - ID of the worker process. - num_workers : int - Total number of workers that will be initialized. - rank : int - The rank of the current process. - seed : int - A random seed used determine the worker seed. - """ - worker_seed = num_workers * rank + worker_id + seed - torch.manual_seed(worker_seed) - random.seed(worker_seed) - - -def setup() -> None: - """Initialize the process group.""" - dist.init_process_group("nccl") - - -def cleanup() -> None: - """Clean up the process group after training.""" - dist.destroy_process_group() - - -def main(): - args = parser.parse_args() - current_time = datetime.now().strftime("%Y-%m-%d-%H-%M") - checkpoint_subdir = ( - f"{args.experiment}_{current_time}" if args.experiment else f"{current_time}" - ) - args.checkpoint_dir = os.path.join(args.checkpoint_dir, checkpoint_subdir) - os.makedirs(args.checkpoint_dir, exist_ok=True) - - print(args) - - # torch.multiprocessing.set_start_method("spawn") - # torch.multiprocessing.set_start_method("spawn") - if args.distributed_mode: - # dist_utils.init_distributed_mode( - # launcher=args.distributed_launcher, - # backend=args.distributed_backend, - # ) - setup() - torch.cuda.set_device(int(os.environ["LOCAL_RANK"])) - torch.cuda.empty_cache() - device_id = torch.cuda.current_device() - else: - device_id = None - - # Data loading. - if args.use_synthetic_data: - print( - f"Using synthetic data for training at {args.synthetic_data_dir} between indices {args.synthetic_index_min} and {args.synthetic_index_max}." - ) - train_dataset = loader.ImageNetSynthetic( - args.data_dir, - args.synthetic_data_dir, - index_min=args.synthetic_index_min, - index_max=args.synthetic_index_max, - generative_augmentation_prob=args.generative_augmentation_prob, - ) - else: - print(f"Using real data for training at {args.data_dir}.") - train_data_dir = os.path.join(args.data_dir, "train") - train_dataset = datasets.ImageFolder(train_data_dir, loader.TwoCropsTransform()) - - train_sampler = None - if dist_utils.is_dist_avail_and_initialized() and args.distributed_mode: - train_sampler = DistributedSampler( - train_dataset, - seed=args.seed, - drop_last=True, - ) - init_fn = partial( - worker_init_fn, - num_workers=args.num_workers, - rank=dist_utils.get_rank(), - seed=args.seed, - ) - - train_loader = torch.utils.data.DataLoader( - train_dataset, - batch_size=args.batch_size, - shuffle=(train_sampler is None), - sampler=train_sampler, - num_workers=args.num_workers, - worker_init_fn=init_fn, - pin_memory=False, - drop_last=True, - ) - if dist_utils.get_rank() == 0: - print(f"Creating model {args.arch}") - model = builder.SimSiam(models.__dict__[args.arch], args.dim, args.pred_dim) - - if args.distributed_mode and dist_utils.is_dist_avail_and_initialized(): - # Apply SyncBN - model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model) - # set the single device scope, otherwise DistributedDataParallel will - # use all available devices - # torch.cuda.set_device(device_id) - model = model.cuda(device_id) - model = DDP(model, device_ids=[device_id]) - else: - raise NotImplementedError("Only DistributedDataParallel is supported.") - if dist_utils.get_rank() == 0: - print(model) # print model after SyncBatchNorm - - # define loss function (criterion) and optimizer - criterion = nn.CosineSimilarity(dim=1).cuda(device_id) - - if args.fix_pred_lr: - optim_params = [ - {"params": model.module.encoder.parameters(), "fix_lr": False}, - {"params": model.module.predictor.parameters(), "fix_lr": True}, - ] - else: - optim_params = model.parameters() - - # infer learning rate before changing batch size - # init_lr = args.lr * args.batch_size / 256.0 - # TODO Hard-code init-lr to match the original paper with bs=512. - init_lr = args.lr * 2.0 - - optimizer = torch.optim.SGD( - optim_params, - init_lr, - momentum=args.momentum, - weight_decay=args.weight_decay, - ) - - start_epoch = 0 - # Optionally resume from a checkpoint - if args.resume_from_checkpoint: - if os.path.isfile(args.resume_from_checkpoint): - print(f"Loading checkpoint: {args.resume_from_checkpoint}") - checkpoint = torch.load(args.resume_from_checkpoint) - start_epoch = checkpoint["epoch"] + 1 - model.load_state_dict(checkpoint["state_dict"]) - optimizer.load_state_dict(checkpoint["optimizer"]) - print(f"Loaded checkpoint {args.resume_from_checkpoint} successfully.") - else: - raise ValueError(f"No checkpoint found at: {args.resume_from_checkpoint}") - - cudnn.benchmark = True - - for epoch in range(start_epoch, args.epochs): - print(f"Starting training epoch: {epoch}") - if dist_utils.is_dist_avail_and_initialized(): - train_sampler.set_epoch(epoch) - adjust_learning_rate(optimizer, init_lr, epoch, args) - - # train for one epoch - train(train_loader, model, criterion, optimizer, epoch, device_id, args) - - # Checkpointing. - if dist_utils.get_rank() == 0: - checkpoint_name = "checkpoint_{:04d}.pth.tar".format(epoch) - checkpoint_file = os.path.join(args.checkpoint_dir, checkpoint_name) - save_checkpoint( - { - "epoch": epoch, - "arch": args.arch, - "state_dict": model.state_dict(), - "optimizer": optimizer.state_dict(), - }, - filename=checkpoint_file, - ) - - -def train(train_loader, model, criterion, optimizer, epoch, device_id, args): - """Single epoch training code.""" - losses = AverageMeter("Loss", ":.4f") - progress = ProgressMeter( - len(train_loader), - [losses], - prefix="Epoch: [{}]".format(epoch), - ) - - # switch to train mode - model.train() - - for i, (images, _) in enumerate(train_loader): - # for images, _ in tqdm(train_loader): - images[0] = images[0].cuda(device_id, non_blocking=True) - images[1] = images[1].cuda(device_id, non_blocking=True) - - # compute output and loss - p1, p2, z1, z2 = model(x1=images[0], x2=images[1]) - loss = -(criterion(p1, z2).mean() + criterion(p2, z1).mean()) * 0.5 - - losses.update(loss.item(), images[0].size(0)) - - # compute gradient and do SGD step - optimizer.zero_grad() - loss.backward() - optimizer.step() - - if i % args.print_freq == 0: - progress.display(i) - - -def save_checkpoint(state, filename="checkpoint.pth.tar"): - """Save state dictionary into a model checkpoint.""" - print(f"Saving checkpoint at: {filename}") - torch.save(state, filename) - - -class AverageMeter(object): - """Computes and stores the average and current value""" - - def __init__(self, name, fmt=":f"): - self.name = name - self.fmt = fmt - self.reset() - - def reset(self): - self.val = 0 - self.avg = 0 - self.sum = 0 - self.count = 0 - - def update(self, val, n=1): - self.val = val - self.sum += val * n - self.count += n - self.avg = self.sum / self.count - - def __str__(self): - fmtstr = "{name} {val" + self.fmt + "} ({avg" + self.fmt + "})" - return fmtstr.format(**self.__dict__) - - -class ProgressMeter(object): - def __init__(self, num_batches, meters, prefix=""): - self.batch_fmtstr = self._get_batch_fmtstr(num_batches) - self.meters = meters - self.prefix = prefix - - def display(self, batch): - entries = [self.prefix + self.batch_fmtstr.format(batch)] - entries += [str(meter) for meter in self.meters] - print("\t".join(entries)) - - def _get_batch_fmtstr(self, num_batches): - num_digits = len(str(num_batches // 1)) - fmt = "{:" + str(num_digits) + "d}" - return "[" + fmt + "/" + fmt.format(num_batches) + "]" - - -def adjust_learning_rate(optimizer, init_lr, epoch, args): - """Decay the learning rate based on schedule.""" - cur_lr = init_lr * 0.5 * (1.0 + math.cos(math.pi * epoch / args.epochs)) - for param_group in optimizer.param_groups: - if "fix_lr" in param_group and param_group["fix_lr"]: - param_group["lr"] = init_lr - else: - param_group["lr"] = cur_lr - - -if __name__ == "__main__": - main() From e7eb3b4b5e8f40b4beccd565717529c4d3d0ded4 Mon Sep 17 00:00:00 2001 From: sanaAyrml Date: Thu, 16 Jan 2025 07:05:48 -0500 Subject: [PATCH 3/7] Update solo library --- solo-learn/solo/args/linear.py | 8 +- .../solo/data/classification_dataloader.py | 109 +++++++++++++++++- solo-learn/solo/data/dali_dataloader.py | 73 +++++++++--- solo-learn/solo/data/inatural_dataset.py | 77 +++++++++++++ 4 files changed, 252 insertions(+), 15 deletions(-) create mode 100644 solo-learn/solo/data/inatural_dataset.py diff --git a/solo-learn/solo/args/linear.py b/solo-learn/solo/args/linear.py index 35330de..b498d7a 100644 --- a/solo-learn/solo/args/linear.py +++ b/solo-learn/solo/args/linear.py @@ -21,6 +21,9 @@ "stl10": 10, "imagenet": 1000, "imagenet100": 100, + "food101": 101, + "places365": 434, + "inaturalist": 8142, } @@ -30,6 +33,9 @@ "stl10", "imagenet", "imagenet100", + "food101", + "places365", + "inaturalist", "custom", ] @@ -163,7 +169,7 @@ def parse_cfg(cfg: omegaconf.DictConfig): ) if cfg.data.format == "dali": - assert cfg.data.dataset in ["imagenet100", "imagenet", "custom"] + assert cfg.data.dataset in ["imagenet100", "imagenet", "inaturalist", "places365", "custom"] # adjust lr according to batch size cfg.num_nodes = omegaconf_select(cfg, "num_nodes", 1) diff --git a/solo-learn/solo/data/classification_dataloader.py b/solo-learn/solo/data/classification_dataloader.py index fc44333..b8fd07d 100644 --- a/solo-learn/solo/data/classification_dataloader.py +++ b/solo-learn/solo/data/classification_dataloader.py @@ -27,7 +27,8 @@ from torch import nn from torch.utils.data import DataLoader, Dataset from torchvision import transforms -from torchvision.datasets import STL10, ImageFolder +from torchvision.datasets import STL10, ImageFolder, Food101, Places365 +from solo.data.inatural_dataset import INAT try: from solo.data.h5_dataset import H5Dataset @@ -136,12 +137,81 @@ def prepare_transforms(dataset: str) -> Tuple[nn.Module, nn.Module]: ), } + food_pipeline = { + "T_train": transforms.Compose( + [ + transforms.RandomResizedCrop(size=224, scale=(0.08, 1.0)), + transforms.RandomHorizontalFlip(), + transforms.ToTensor(), + transforms.Normalize( + mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD + ), + ] + ), + "T_val": transforms.Compose( + [ + transforms.Resize((224, 224)), + transforms.ToTensor(), + transforms.Normalize( + mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD + ), + ] + ), + } + + place_pipeline = { + "T_train": transforms.Compose( + [ + transforms.RandomResizedCrop(size=224, scale=(0.08, 1.0)), + transforms.RandomHorizontalFlip(), + transforms.ToTensor(), + transforms.Normalize( + mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD + ), + ] + ), + "T_val": transforms.Compose( + [ + transforms.Resize((224, 224)), + transforms.ToTensor(), + transforms.Normalize( + mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD + ), + ] + ), + } + + inat_pipeline = { + "T_train": transforms.Compose( + [ + transforms.RandomResizedCrop(size=224, scale=(0.08, 1.0)), + transforms.RandomHorizontalFlip(), + transforms.ToTensor(), + transforms.Normalize( + mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD + ), + ] + ), + "T_val": transforms.Compose( + [ + transforms.Resize((224, 224)), + transforms.ToTensor(), + transforms.Normalize( + mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD + ), + ] + ), + } + custom_pipeline = build_custom_pipeline() pipelines = { "cifar10": cifar_pipeline, "cifar100": cifar_pipeline, "stl10": stl_pipeline, + "food101": food_pipeline, + "places365": place_pipeline, + "inaturalist": inat_pipeline, "imagenet100": imagenet_pipeline, "imagenet": imagenet_pipeline, "custom": custom_pipeline, @@ -201,6 +271,9 @@ def prepare_datasets( "cifar10", "cifar100", "stl10", + "food101", + "places365", + "inaturalist", "imagenet", "imagenet100", "custom", @@ -221,7 +294,41 @@ def prepare_datasets( download=download, transform=T_val, ) + elif dataset == "food101": + print("=> using food101 dataset.", flush=True) + train_dataset = Food101( + root=train_data_path, + split="train", + transform=T_train, + ) + val_dataset = Food101( + root=val_data_path, + split="test", + transform=T_val, + ) + elif dataset == "places365": + train_dataset = Places365( + root=train_data_path, + split="train-standard", + transform=T_train, + ) + val_dataset = Places365( + root=val_data_path, + split="val", + transform=T_val, + ) + elif dataset == "inaturalist": + train_dataset = INAT( + root=train_data_path, + ann_file=os.path.join(train_data_path, "train2018.json"), + transform=T_train, + ) + val_dataset = INAT( + root=val_data_path, + ann_file=os.path.join(val_data_path, "val2018.json"), + transform=T_val, + ) elif dataset == "stl10": train_dataset = STL10( train_data_path, diff --git a/solo-learn/solo/data/dali_dataloader.py b/solo-learn/solo/data/dali_dataloader.py index 8451846..406b56b 100644 --- a/solo-learn/solo/data/dali_dataloader.py +++ b/solo-learn/solo/data/dali_dataloader.py @@ -20,7 +20,7 @@ import os import random from pathlib import Path -from typing import Callable, List, Optional, Union +from typing import Callable, List, Optional, Union, Tuple import lightning.pytorch as pl import nvidia.dali.fn as fn @@ -32,6 +32,7 @@ from nvidia.dali import pipeline_def from nvidia.dali.plugin.pytorch import DALIGenericIterator, LastBatchPolicy from timm.data.constants import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD +import json from solo.data.temp_dali_fix import TempDALIGenericIterator from solo.utils.misc import omegaconf_select @@ -202,6 +203,7 @@ def __call__(self, images): class NormalPipelineBuilder: def __init__( self, + dataset: str, data_path: str, batch_size: int, device: str, @@ -220,6 +222,7 @@ def __init__( are normalized. Args: + dataset (str): dataset name. data_path (str): directory that contains the data. batch_size (int): batch size. device (str): device on which the operation will be performed. @@ -244,17 +247,59 @@ def __init__( self.device = device self.validation = validation - - # manually load files and labels - labels = sorted( - Path(entry.name) for entry in os.scandir(data_path) if entry.is_dir() - ) - data = [ - (data_path / label / file, label_idx) - for label_idx, label in enumerate(labels) - for file in sorted(os.listdir(data_path / label)) - ] - files, labels = map(list, zip(*data)) + + if dataset in ["imagenet", "imagenet100"]: + # manually load files and labels + labels = sorted( + Path(entry.name) for entry in os.scandir(data_path) if entry.is_dir() + ) + data = [ + (data_path / label / file, label_idx) + for label_idx, label in enumerate(labels) + for file in sorted(os.listdir(data_path / label)) + ] + files, labels = map(list, zip(*data)) + elif dataset == "places365": + if not validation: + split = "train-standard" + else: + split = "val" + _FILE_LIST_META = { + "train-standard": ("places365_train_standard.txt", "30f37515461640559006b8329efbed1a", "data_large_standard"), + "train-challenge": ("places365_train_challenge.txt", "b2931dc997b8c33c27e7329c073a6b57", "data_large"), + "val": ("places365_val.txt", "e9f2fd57bfd9d07630173f4e8708e4b1", "val_large"), + } + def process(line: str, image_dir: str, sep="/") -> Tuple[Path, int]: + image, idx = line.split() + return Path(os.path.join(data_path, image_dir, image.lstrip(sep).replace(sep, os.sep))), int(idx) + + file, md5, image_dir = _FILE_LIST_META[split] + file = os.path.join(data_path, file) + + with open(file) as fh: + data = [process(line, image_dir) for line in fh] + files, labels = map(list, zip(*data)) + print(files[0], labels[0], flush=True) + elif dataset == "inaturalist": + if not validation: + ann_file = os.path.join(data_path, "train2018.json") + else: + ann_file = os.path.join(data_path, "val2018.json") + # load annotations + print("Loading annotations from: " + os.path.basename(ann_file)) + with open(ann_file) as data_file: + ann_data = json.load(data_file) + + # set up the filenames and annotations + files: List[str] = [os.path.join(data_path, aa["file_name"]) for aa in ann_data["images"]] + + # if we dont have class labels set them to '0' + if "annotations" in ann_data.keys(): + labels = [aa["category_id"] for aa in ann_data["annotations"]] + else: + labels= [0] * len(files) + else: + raise NotImplementedError(f"Dataset {dataset} is not supported.") # sample data if needed if data_fraction > 0: @@ -1003,7 +1048,7 @@ def __init__( assert dali_device in ["gpu", "cpu"] # handle custom data by creating the needed pipeline - if dataset in ["imagenet100", "imagenet"]: + if dataset in ["imagenet100", "imagenet", "places365", "inaturalist"]: self.pipeline_class = NormalPipelineBuilder elif dataset == "custom": self.pipeline_class = CustomNormalPipelineBuilder @@ -1040,6 +1085,7 @@ def setup(self, stage: Optional[str] = None): self.device = torch.device("cpu") train_pipeline_builder = self.pipeline_class( + self.dataset, self.train_data_path, validation=False, batch_size=self.batch_size, @@ -1068,6 +1114,7 @@ def setup(self, stage: Optional[str] = None): ) val_pipeline_builder = self.pipeline_class( + self.dataset, self.val_data_path, validation=True, batch_size=self.batch_size, diff --git a/solo-learn/solo/data/inatural_dataset.py b/solo-learn/solo/data/inatural_dataset.py new file mode 100644 index 0000000..65b689f --- /dev/null +++ b/solo-learn/solo/data/inatural_dataset.py @@ -0,0 +1,77 @@ +import torch.utils.data as data +from PIL import Image +import os +import json +from torchvision import transforms +import random +import numpy as np + + +def default_loader(path): + return Image.open(path).convert("RGB") + + +def load_taxonomy(ann_data, tax_levels, classes): + # loads the taxonomy data and converts to ints + taxonomy = {} + + if "categories" in ann_data.keys(): + num_classes = len(ann_data["categories"]) + for tt in tax_levels: + tax_data = [aa[tt] for aa in ann_data["categories"]] + _, tax_id = np.unique(tax_data, return_inverse=True) + taxonomy[tt] = dict(zip(range(num_classes), list(tax_id))) + else: + # set up dummy data + for tt in tax_levels: + taxonomy[tt] = dict(zip([0], [0])) + + # create a dictionary of lists containing taxonomic labels + classes_taxonomic = {} + for cc in np.unique(classes): + tax_ids = [0] * len(tax_levels) + for ii, tt in enumerate(tax_levels): + tax_ids[ii] = taxonomy[tt][cc] + classes_taxonomic[cc] = tax_ids + + return taxonomy, classes_taxonomic + + +class INAT(data.Dataset): + def __init__(self, root, ann_file, transform): + # load annotations + print("Loading annotations from: " + os.path.basename(ann_file)) + with open(ann_file) as data_file: + ann_data = json.load(data_file) + + # set up the filenames and annotations + self.imgs = [aa["file_name"] for aa in ann_data["images"]] + self.ids = [aa["id"] for aa in ann_data["images"]] + + # if we dont have class labels set them to '0' + if "annotations" in ann_data.keys(): + self.classes = [aa["category_id"] for aa in ann_data["annotations"]] + else: + self.classes = [0] * len(self.imgs) + + # print out some stats + print("\t" + str(len(self.imgs)) + " images") + print("\t" + str(len(set(self.classes))) + " classes") + + self.root = root + self.loader = default_loader + + # augmentation params + self.transform = transform + + def __getitem__(self, index): + path = self.root + self.imgs[index] + img = self.loader(path) + species_id = self.classes[index] + + img = self.transform(img) + + return img, species_id + + def __len__(self): + return len(self.imgs) From 340e6ebc6283be5c9cc1bef2a6e4cd679a4ca764 Mon Sep 17 00:00:00 2001 From: sanaAyrml Date: Thu, 16 Jan 2025 07:07:19 -0500 Subject: [PATCH 4/7] Delete extra scripts anf keep needed scripts --- .../scripts/finetune/imagenet-100/mae.yaml | 52 ------------ .../finetune/imagenet-100/wandb/mhug.yaml | 3 - .../finetune/imagenet-100/wandb/private.yaml | 3 - solo-learn/scripts/finetune/imagenet/mae.yaml | 52 ------------ .../scripts/finetune/imagenet/wandb/mhug.yaml | 3 - .../finetune/imagenet/wandb/private.yaml | 3 - solo-learn/scripts/knn/imagenet-100/knn.sh | 11 --- .../{imagenet-100 => cifar10}/barlow.yaml | 25 +++--- .../scripts/linear/cifar10/barlow_diff.yaml | 46 ++++++++++ .../scripts/linear/cifar10/barlow_icgan.yaml | 45 ++++++++++ .../ressl.yaml => cifar10/byol.yaml} | 32 +++---- .../scripts/linear/cifar10/byol_diff.yaml | 46 ++++++++++ .../scripts/linear/cifar10/byol_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/cifar10/moco.yaml | 45 ++++++++++ .../scripts/linear/cifar10/moco_diff.yaml | 45 ++++++++++ .../scripts/linear/cifar10/moco_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/cifar10/simclr.yaml | 46 ++++++++++ .../scripts/linear/cifar10/simclr_diff.yaml | 46 ++++++++++ .../scripts/linear/cifar10/simclr_icgan.yaml | 46 ++++++++++ .../scripts/linear/cifar10/simsiam.yaml | 46 ++++++++++ .../scripts/linear/cifar10/simsiam_diff.yaml | 46 ++++++++++ .../scripts/linear/cifar10/simsiam_icgan.yaml | 46 ++++++++++ .../scripts/linear/cifar100/barlow.yaml | 46 ++++++++++ .../scripts/linear/cifar100/barlow_diff.yaml | 46 ++++++++++ .../scripts/linear/cifar100/barlow_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/cifar100/byol.yaml | 46 ++++++++++ .../scripts/linear/cifar100/byol_diff.yaml | 46 ++++++++++ .../scripts/linear/cifar100/byol_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/cifar100/moco.yaml | 46 ++++++++++ .../scripts/linear/cifar100/moco_diff.yaml | 46 ++++++++++ .../scripts/linear/cifar100/moco_icgan.yaml | 46 ++++++++++ .../scripts/linear/cifar100/simclr.yaml | 46 ++++++++++ .../scripts/linear/cifar100/simclr_diff.yaml | 46 ++++++++++ .../scripts/linear/cifar100/simclr_icgan.yaml | 46 ++++++++++ .../scripts/linear/cifar100/simsiam.yaml | 46 ++++++++++ .../scripts/linear/cifar100/simsiam_diff.yaml | 46 ++++++++++ .../linear/cifar100/simsiam_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/food/barlow.yaml | 46 ++++++++++ .../scripts/linear/food/barlow_diff.yaml | 46 ++++++++++ .../scripts/linear/food/barlow_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/food/byol.yaml | 46 ++++++++++ solo-learn/scripts/linear/food/byol_diff.yaml | 46 ++++++++++ .../scripts/linear/food/byol_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/food/moco.yaml | 46 ++++++++++ solo-learn/scripts/linear/food/moco_diff.yaml | 46 ++++++++++ .../scripts/linear/food/moco_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/food/simclr.yaml | 46 ++++++++++ .../scripts/linear/food/simclr_diff.yaml | 46 ++++++++++ .../scripts/linear/food/simclr_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/food/simsiam.yaml | 46 ++++++++++ .../scripts/linear/food/simsiam_diff.yaml | 46 ++++++++++ .../scripts/linear/food/simsiam_icgan.yaml | 46 ++++++++++ .../linear/imagenet-100/deepclusterv2.yaml | 45 ---------- .../linear/imagenet-100/mocov3_vit.yaml | 45 ---------- .../scripts/linear/imagenet-100/vibcreg.yaml | 45 ---------- .../scripts/linear/imagenet-100/vicreg.yaml | 45 ---------- .../linear/imagenet-100/wandb/mhug.yaml | 3 - .../linear/imagenet-100/wandb/private.yaml | 3 - .../scripts/linear/imagenet/barlow.yaml | 25 +++--- .../dino.yaml => imagenet/barlow_diff.yaml} | 29 +++---- .../swav.yaml => imagenet/barlow_icgan.yaml} | 29 +++---- solo-learn/scripts/linear/imagenet/byol.yaml | 19 +++-- .../nnclr.yaml => imagenet/byol_diff.yaml} | 29 +++---- .../scripts/linear/imagenet/byol_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/imagenet/moco.yaml | 45 ++++++++++ .../moco_diff.yaml} | 27 +++--- .../scripts/linear/imagenet/moco_icgan.yaml | 46 ++++++++++ .../scripts/linear/imagenet/mocov2plus.yaml | 45 ---------- .../{imagenet-100 => imagenet}/simclr.yaml | 27 +++--- .../scripts/linear/imagenet/simclr_diff.yaml | 46 ++++++++++ .../scripts/linear/imagenet/simclr_icgan.yaml | 46 ++++++++++ .../scripts/linear/imagenet/simsiam.yaml | 23 ++--- .../scripts/linear/imagenet/simsiam_diff.yaml | 46 ++++++++++ .../linear/imagenet/simsiam_icgan.yaml | 46 ++++++++++ .../scripts/linear/inaturalist/barlow.yaml | 46 ++++++++++ .../linear/inaturalist/barlow_diff.yaml | 46 ++++++++++ .../barlow_icgan.yaml} | 29 +++---- .../{imagenet-100 => inaturalist}/byol.yaml | 27 +++--- .../scripts/linear/inaturalist/byol_diff.yaml | 46 ++++++++++ .../linear/inaturalist/byol_icgan.yaml | 46 ++++++++++ .../scripts/linear/inaturalist/moco.yaml | 45 ++++++++++ .../scripts/linear/inaturalist/moco_diff.yaml | 46 ++++++++++ .../linear/inaturalist/moco_icgan.yaml | 46 ++++++++++ .../scripts/linear/inaturalist/simclr.yaml | 46 ++++++++++ .../linear/inaturalist/simclr_diff.yaml | 46 ++++++++++ .../linear/inaturalist/simclr_icgan.yaml | 46 ++++++++++ .../simsiam.yaml | 27 +++--- .../linear/inaturalist/simsiam_diff.yaml | 46 ++++++++++ .../linear/inaturalist/simsiam_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/places/barlow.yaml | 46 ++++++++++ .../scripts/linear/places/barlow_diff.yaml | 46 ++++++++++ .../scripts/linear/places/barlow_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/places/byol.yaml | 46 ++++++++++ .../scripts/linear/places/byol_diff.yaml | 46 ++++++++++ .../scripts/linear/places/byol_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/places/moco.yaml | 45 ++++++++++ .../scripts/linear/places/moco_diff.yaml | 46 ++++++++++ .../scripts/linear/places/moco_icgan.yaml | 45 ++++++++++ solo-learn/scripts/linear/places/simclr.yaml | 46 ++++++++++ .../scripts/linear/places/simclr_diff.yaml | 46 ++++++++++ .../scripts/linear/places/simclr_icgan.yaml | 46 ++++++++++ solo-learn/scripts/linear/places/simsiam.yaml | 46 ++++++++++ .../scripts/linear/places/simsiam_diff.yaml | 46 ++++++++++ .../scripts/linear/places/simsiam_icgan.yaml | 46 ++++++++++ .../cifar-multicrop/augmentations/swav.yaml | 68 --------------- .../pretrain/cifar-multicrop/swav.yaml | 58 ------------- .../pretrain/cifar-multicrop/wandb/mhug.yaml | 3 - .../cifar-multicrop/wandb/private.yaml | 3 - .../scripts/pretrain/cifar/all4one.yaml | 58 ------------- .../cifar/augmentations/asymmetric.yaml | 45 ---------- .../cifar/augmentations/reconstruction.yaml | 18 ---- .../pretrain/cifar/augmentations/ressl.yaml | 41 --------- .../cifar/augmentations/symmetric.yaml | 22 ----- .../cifar/augmentations/symmetric_weak.yaml | 22 ----- solo-learn/scripts/pretrain/cifar/barlow.yaml | 53 ------------ .../scripts/pretrain/cifar/deepclusterv2.yaml | 56 ------------- solo-learn/scripts/pretrain/cifar/dino.yaml | 56 ------------- solo-learn/scripts/pretrain/cifar/mae.yaml | 56 ------------- solo-learn/scripts/pretrain/cifar/nnclr.yaml | 55 ------------ solo-learn/scripts/pretrain/cifar/nnsiam.yaml | 53 ------------ solo-learn/scripts/pretrain/cifar/ressl.yaml | 56 ------------- solo-learn/scripts/pretrain/cifar/simclr.yaml | 53 ------------ .../scripts/pretrain/cifar/simsiam.yaml | 50 ----------- solo-learn/scripts/pretrain/cifar/supcon.yaml | 49 ----------- solo-learn/scripts/pretrain/cifar/swav.yaml | 57 ------------- .../scripts/pretrain/cifar/vibcreg.yaml | 77 ----------------- solo-learn/scripts/pretrain/cifar/vicreg.yaml | 83 ------------------ .../scripts/pretrain/cifar/wandb/mhug.yaml | 3 - .../scripts/pretrain/cifar/wandb/private.yaml | 3 - solo-learn/scripts/pretrain/cifar/wmse.yaml | 73 ---------------- .../custom/augmentations/asymmetric.yaml | 45 ---------- .../custom/augmentations/symmetric.yaml | 22 ----- solo-learn/scripts/pretrain/custom/byol.yaml | 63 -------------- .../scripts/pretrain/custom/wandb/mhug.yaml | 3 - .../pretrain/custom/wandb/private.yaml | 3 - .../augmentations/asymmetric.yaml | 68 --------------- .../augmentations/symmetric.yaml | 45 ---------- .../imagenet-100-multicrop/supcon.yaml | 49 ----------- .../imagenet-100-multicrop/wandb/mhug.yaml | 3 - .../imagenet-100-multicrop/wandb/private.yaml | 3 - .../scripts/pretrain/imagenet-100/all4one.yml | 55 ------------ .../augmentations/reconstruction.yaml | 18 ---- .../imagenet-100/augmentations/ressl.yaml | 41 --------- .../imagenet-100/augmentations/symmetric.yaml | 22 ----- .../augmentations/symmetric_weak.yaml | 22 ----- .../imagenet-100/augmentations/vicreg.yaml | 22 ----- .../imagenet-100/augmentations/wmse.yaml | 45 ---------- .../scripts/pretrain/imagenet-100/barlow.yaml | 53 ------------ .../pretrain/imagenet-100/deepclusterv2.yaml | 58 ------------- .../scripts/pretrain/imagenet-100/dino.yaml | 57 ------------- .../pretrain/imagenet-100/dino_vit.yaml | 54 ------------ .../scripts/pretrain/imagenet-100/mae.yaml | 54 ------------ .../scripts/pretrain/imagenet-100/mocov3.yaml | 57 ------------- .../pretrain/imagenet-100/mocov3_vit.yaml | 53 ------------ .../scripts/pretrain/imagenet-100/nnclr.yaml | 55 ------------ .../scripts/pretrain/imagenet-100/ressl.yaml | 56 ------------- .../pretrain/imagenet-100/simsiam.yaml | 51 ----------- .../scripts/pretrain/imagenet-100/supcon.yaml | 49 ----------- .../scripts/pretrain/imagenet-100/swav.yaml | 57 ------------- .../pretrain/imagenet-100/vibcreg.yaml | 56 ------------- .../scripts/pretrain/imagenet-100/vicreg.yaml | 84 ------------------- .../pretrain/imagenet-100/wandb/mhug.yaml | 3 - .../pretrain/imagenet-100/wandb/private.yaml | 3 - .../scripts/pretrain/imagenet-100/wmse.yaml | 50 ----------- .../augmentations/synthetic_symmetric.yaml | 2 +- .../synthetic_symmetric_weak.yaml} | 13 +-- .../scripts/pretrain/imagenet/barlow.yaml | 10 +-- .../byol.yaml => imagenet/barlow_diff.yaml} | 42 +++++----- .../byol.yaml => imagenet/barlow_icgan.yaml} | 40 ++++----- .../scripts/pretrain/imagenet/byol.yaml | 12 +-- .../nnbyol.yaml => imagenet/byol_diff.yaml} | 38 +++++---- .../byol.yaml => imagenet/byol_icgan.yaml} | 33 ++++---- solo-learn/scripts/pretrain/imagenet/mae.yaml | 57 ------------- .../imagenet/{mocov2plus.yaml => moco.yaml} | 10 +-- .../moco_diff.yaml} | 34 ++++---- .../moco_icgan.yaml} | 30 ++++--- .../scripts/pretrain/imagenet/simclr.yaml | 10 +-- .../simclr.yaml => imagenet/simclr_diff.yaml} | 26 +++--- .../simclr_icgan.yaml} | 26 +++--- .../simclr_paper_synth_stable_25.yaml} | 34 ++++---- ...yaml => simclr_paper_synth_stable_50.yaml} | 14 ++-- .../simclr_paper_synth_stable_75.yaml | 57 +++++++++++++ .../scripts/pretrain/imagenet/simsiam.yaml | 27 +++--- .../pretrain/imagenet/simsiam_diff.yaml | 51 +++++++++++ .../pretrain/imagenet/simsiam_icgan.yaml | 51 +++++++++++ solo-learn/scripts/umap/imagenet-100/umap.sh | 7 -- 186 files changed, 4089 insertions(+), 3326 deletions(-) delete mode 100644 solo-learn/scripts/finetune/imagenet-100/mae.yaml delete mode 100644 solo-learn/scripts/finetune/imagenet-100/wandb/mhug.yaml delete mode 100644 solo-learn/scripts/finetune/imagenet-100/wandb/private.yaml delete mode 100644 solo-learn/scripts/finetune/imagenet/mae.yaml delete mode 100644 solo-learn/scripts/finetune/imagenet/wandb/mhug.yaml delete mode 100644 solo-learn/scripts/finetune/imagenet/wandb/private.yaml delete mode 100644 solo-learn/scripts/knn/imagenet-100/knn.sh rename solo-learn/scripts/linear/{imagenet-100 => cifar10}/barlow.yaml (58%) create mode 100644 solo-learn/scripts/linear/cifar10/barlow_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar10/barlow_icgan.yaml rename solo-learn/scripts/linear/{imagenet-100/ressl.yaml => cifar10/byol.yaml} (54%) create mode 100644 solo-learn/scripts/linear/cifar10/byol_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar10/byol_icgan.yaml create mode 100644 solo-learn/scripts/linear/cifar10/moco.yaml create mode 100644 solo-learn/scripts/linear/cifar10/moco_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar10/moco_icgan.yaml create mode 100644 solo-learn/scripts/linear/cifar10/simclr.yaml create mode 100644 solo-learn/scripts/linear/cifar10/simclr_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar10/simclr_icgan.yaml create mode 100644 solo-learn/scripts/linear/cifar10/simsiam.yaml create mode 100644 solo-learn/scripts/linear/cifar10/simsiam_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar10/simsiam_icgan.yaml create mode 100644 solo-learn/scripts/linear/cifar100/barlow.yaml create mode 100644 solo-learn/scripts/linear/cifar100/barlow_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar100/barlow_icgan.yaml create mode 100644 solo-learn/scripts/linear/cifar100/byol.yaml create mode 100644 solo-learn/scripts/linear/cifar100/byol_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar100/byol_icgan.yaml create mode 100644 solo-learn/scripts/linear/cifar100/moco.yaml create mode 100644 solo-learn/scripts/linear/cifar100/moco_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar100/moco_icgan.yaml create mode 100644 solo-learn/scripts/linear/cifar100/simclr.yaml create mode 100644 solo-learn/scripts/linear/cifar100/simclr_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar100/simclr_icgan.yaml create mode 100644 solo-learn/scripts/linear/cifar100/simsiam.yaml create mode 100644 solo-learn/scripts/linear/cifar100/simsiam_diff.yaml create mode 100644 solo-learn/scripts/linear/cifar100/simsiam_icgan.yaml create mode 100644 solo-learn/scripts/linear/food/barlow.yaml create mode 100644 solo-learn/scripts/linear/food/barlow_diff.yaml create mode 100644 solo-learn/scripts/linear/food/barlow_icgan.yaml create mode 100644 solo-learn/scripts/linear/food/byol.yaml create mode 100644 solo-learn/scripts/linear/food/byol_diff.yaml create mode 100644 solo-learn/scripts/linear/food/byol_icgan.yaml create mode 100644 solo-learn/scripts/linear/food/moco.yaml create mode 100644 solo-learn/scripts/linear/food/moco_diff.yaml create mode 100644 solo-learn/scripts/linear/food/moco_icgan.yaml create mode 100644 solo-learn/scripts/linear/food/simclr.yaml create mode 100644 solo-learn/scripts/linear/food/simclr_diff.yaml create mode 100644 solo-learn/scripts/linear/food/simclr_icgan.yaml create mode 100644 solo-learn/scripts/linear/food/simsiam.yaml create mode 100644 solo-learn/scripts/linear/food/simsiam_diff.yaml create mode 100644 solo-learn/scripts/linear/food/simsiam_icgan.yaml delete mode 100644 solo-learn/scripts/linear/imagenet-100/deepclusterv2.yaml delete mode 100644 solo-learn/scripts/linear/imagenet-100/mocov3_vit.yaml delete mode 100644 solo-learn/scripts/linear/imagenet-100/vibcreg.yaml delete mode 100644 solo-learn/scripts/linear/imagenet-100/vicreg.yaml delete mode 100644 solo-learn/scripts/linear/imagenet-100/wandb/mhug.yaml delete mode 100644 solo-learn/scripts/linear/imagenet-100/wandb/private.yaml rename solo-learn/scripts/linear/{imagenet-100/dino.yaml => imagenet/barlow_diff.yaml} (57%) rename solo-learn/scripts/linear/{imagenet-100/swav.yaml => imagenet/barlow_icgan.yaml} (57%) rename solo-learn/scripts/linear/{imagenet-100/nnclr.yaml => imagenet/byol_diff.yaml} (57%) create mode 100644 solo-learn/scripts/linear/imagenet/byol_icgan.yaml create mode 100644 solo-learn/scripts/linear/imagenet/moco.yaml rename solo-learn/scripts/linear/{imagenet-100/mocov2plus.yaml => imagenet/moco_diff.yaml} (60%) create mode 100644 solo-learn/scripts/linear/imagenet/moco_icgan.yaml delete mode 100644 solo-learn/scripts/linear/imagenet/mocov2plus.yaml rename solo-learn/scripts/linear/{imagenet-100 => imagenet}/simclr.yaml (60%) create mode 100644 solo-learn/scripts/linear/imagenet/simclr_diff.yaml create mode 100644 solo-learn/scripts/linear/imagenet/simclr_icgan.yaml create mode 100644 solo-learn/scripts/linear/imagenet/simsiam_diff.yaml create mode 100644 solo-learn/scripts/linear/imagenet/simsiam_icgan.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/barlow.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/barlow_diff.yaml rename solo-learn/scripts/linear/{imagenet-100/mocov3.yaml => inaturalist/barlow_icgan.yaml} (57%) rename solo-learn/scripts/linear/{imagenet-100 => inaturalist}/byol.yaml (60%) create mode 100644 solo-learn/scripts/linear/inaturalist/byol_diff.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/byol_icgan.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/moco.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/moco_diff.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/moco_icgan.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/simclr.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/simclr_diff.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/simclr_icgan.yaml rename solo-learn/scripts/linear/{imagenet-100 => inaturalist}/simsiam.yaml (60%) create mode 100644 solo-learn/scripts/linear/inaturalist/simsiam_diff.yaml create mode 100644 solo-learn/scripts/linear/inaturalist/simsiam_icgan.yaml create mode 100644 solo-learn/scripts/linear/places/barlow.yaml create mode 100644 solo-learn/scripts/linear/places/barlow_diff.yaml create mode 100644 solo-learn/scripts/linear/places/barlow_icgan.yaml create mode 100644 solo-learn/scripts/linear/places/byol.yaml create mode 100644 solo-learn/scripts/linear/places/byol_diff.yaml create mode 100644 solo-learn/scripts/linear/places/byol_icgan.yaml create mode 100644 solo-learn/scripts/linear/places/moco.yaml create mode 100644 solo-learn/scripts/linear/places/moco_diff.yaml create mode 100644 solo-learn/scripts/linear/places/moco_icgan.yaml create mode 100644 solo-learn/scripts/linear/places/simclr.yaml create mode 100644 solo-learn/scripts/linear/places/simclr_diff.yaml create mode 100644 solo-learn/scripts/linear/places/simclr_icgan.yaml create mode 100644 solo-learn/scripts/linear/places/simsiam.yaml create mode 100644 solo-learn/scripts/linear/places/simsiam_diff.yaml create mode 100644 solo-learn/scripts/linear/places/simsiam_icgan.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar-multicrop/augmentations/swav.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar-multicrop/swav.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar-multicrop/wandb/mhug.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar-multicrop/wandb/private.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/all4one.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/augmentations/asymmetric.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/augmentations/reconstruction.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/augmentations/ressl.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/augmentations/symmetric.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/augmentations/symmetric_weak.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/barlow.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/deepclusterv2.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/dino.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/mae.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/nnclr.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/nnsiam.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/ressl.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/simclr.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/simsiam.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/supcon.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/swav.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/vibcreg.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/vicreg.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/wandb/mhug.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/wandb/private.yaml delete mode 100644 solo-learn/scripts/pretrain/cifar/wmse.yaml delete mode 100644 solo-learn/scripts/pretrain/custom/augmentations/asymmetric.yaml delete mode 100644 solo-learn/scripts/pretrain/custom/augmentations/symmetric.yaml delete mode 100644 solo-learn/scripts/pretrain/custom/byol.yaml delete mode 100644 solo-learn/scripts/pretrain/custom/wandb/mhug.yaml delete mode 100644 solo-learn/scripts/pretrain/custom/wandb/private.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100-multicrop/augmentations/asymmetric.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100-multicrop/augmentations/symmetric.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100-multicrop/supcon.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100-multicrop/wandb/mhug.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100-multicrop/wandb/private.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/all4one.yml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/augmentations/reconstruction.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/augmentations/ressl.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/augmentations/symmetric.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/augmentations/symmetric_weak.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/augmentations/vicreg.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/augmentations/wmse.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/barlow.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/deepclusterv2.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/dino.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/dino_vit.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/mae.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/mocov3.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/mocov3_vit.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/nnclr.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/ressl.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/simsiam.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/supcon.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/swav.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/vibcreg.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/vicreg.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/wandb/mhug.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/wandb/private.yaml delete mode 100644 solo-learn/scripts/pretrain/imagenet-100/wmse.yaml rename solo-learn/scripts/pretrain/{imagenet-100/augmentations/asymmetric.yaml => imagenet/augmentations/synthetic_symmetric_weak.yaml} (66%) rename solo-learn/scripts/pretrain/{imagenet-100/byol.yaml => imagenet/barlow_diff.yaml} (56%) rename solo-learn/scripts/pretrain/{imagenet-100-multicrop/byol.yaml => imagenet/barlow_icgan.yaml} (58%) rename solo-learn/scripts/pretrain/{cifar/nnbyol.yaml => imagenet/byol_diff.yaml} (61%) rename solo-learn/scripts/pretrain/{cifar/byol.yaml => imagenet/byol_icgan.yaml} (63%) delete mode 100644 solo-learn/scripts/pretrain/imagenet/mae.yaml rename solo-learn/scripts/pretrain/imagenet/{mocov2plus.yaml => moco.yaml} (84%) rename solo-learn/scripts/pretrain/{cifar/mocov2plus.yaml => imagenet/moco_diff.yaml} (57%) rename solo-learn/scripts/pretrain/{imagenet-100/mocov2plus.yaml => imagenet/moco_icgan.yaml} (61%) rename solo-learn/scripts/pretrain/{imagenet-100/simclr.yaml => imagenet/simclr_diff.yaml} (65%) rename solo-learn/scripts/pretrain/{imagenet-100-multicrop/simclr.yaml => imagenet/simclr_icgan.yaml} (65%) rename solo-learn/scripts/pretrain/{cifar/mocov3.yaml => imagenet/simclr_paper_synth_stable_25.yaml} (60%) rename solo-learn/scripts/pretrain/imagenet/{simclr_synthetic.yaml => simclr_paper_synth_stable_50.yaml} (73%) create mode 100644 solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_75.yaml create mode 100644 solo-learn/scripts/pretrain/imagenet/simsiam_diff.yaml create mode 100644 solo-learn/scripts/pretrain/imagenet/simsiam_icgan.yaml delete mode 100644 solo-learn/scripts/umap/imagenet-100/umap.sh diff --git a/solo-learn/scripts/finetune/imagenet-100/mae.yaml b/solo-learn/scripts/finetune/imagenet-100/mae.yaml deleted file mode 100644 index a51b725..0000000 --- a/solo-learn/scripts/finetune/imagenet-100/mae.yaml +++ /dev/null @@ -1,52 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mae-imagenet100-finetune" -pretrained_feature_extractor: None -backbone: - name: "vit_base" - kwargs: - drop_path_rate: 0.1 -pretrain_method: "mae" -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "adamw" - batch_size: 64 - lr: 5e-4 - weight_decay: 0.05 - layer_decay: 0.75 -scheduler: - name: "warmup_cosine" - warmup_start_lr: 0.0 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True -label_smoothing: 0.1 -mixup: 0.8 -cutmix: 1.0 -finetune: True - -# overwrite PL stuff -max_epochs: 100 -devices: [0, 1, 2, 3, 4, 5, 6, 7] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16 diff --git a/solo-learn/scripts/finetune/imagenet-100/wandb/mhug.yaml b/solo-learn/scripts/finetune/imagenet-100/wandb/mhug.yaml deleted file mode 100644 index c842e44..0000000 --- a/solo-learn/scripts/finetune/imagenet-100/wandb/mhug.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: unitn-mhug -project: "gen-ssl" diff --git a/solo-learn/scripts/finetune/imagenet-100/wandb/private.yaml b/solo-learn/scripts/finetune/imagenet-100/wandb/private.yaml deleted file mode 100644 index ad4e200..0000000 --- a/solo-learn/scripts/finetune/imagenet-100/wandb/private.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: None -project: "gen-ssl" diff --git a/solo-learn/scripts/finetune/imagenet/mae.yaml b/solo-learn/scripts/finetune/imagenet/mae.yaml deleted file mode 100644 index f3c0453..0000000 --- a/solo-learn/scripts/finetune/imagenet/mae.yaml +++ /dev/null @@ -1,52 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mae-imagenet-finetune" -pretrained_feature_extractor: None -backbone: - name: "vit_base" - kwargs: - drop_path_rate: 0.1 -pretrain_method: "mae" -data: - dataset: "imagenet" - train_path: "./datasets/imagenet/train" - val_path: "./datasets/imagenet/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "adamw" - batch_size: 64 - lr: 5e-4 - weight_decay: 0.05 - layer_decay: 0.75 -scheduler: - name: "warmup_cosine" - warmup_start_lr: 0.0 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True -label_smoothing: 0.1 -mixup: 0.8 -cutmix: 1.0 -finetune: True - -# overwrite PL stuff -max_epochs: 100 -devices: [0, 1, 2, 3, 4, 5, 6, 7] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16 diff --git a/solo-learn/scripts/finetune/imagenet/wandb/mhug.yaml b/solo-learn/scripts/finetune/imagenet/wandb/mhug.yaml deleted file mode 100644 index c842e44..0000000 --- a/solo-learn/scripts/finetune/imagenet/wandb/mhug.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: unitn-mhug -project: "gen-ssl" diff --git a/solo-learn/scripts/finetune/imagenet/wandb/private.yaml b/solo-learn/scripts/finetune/imagenet/wandb/private.yaml deleted file mode 100644 index ad4e200..0000000 --- a/solo-learn/scripts/finetune/imagenet/wandb/private.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: None -project: "gen-ssl" diff --git a/solo-learn/scripts/knn/imagenet-100/knn.sh b/solo-learn/scripts/knn/imagenet-100/knn.sh deleted file mode 100644 index b742093..0000000 --- a/solo-learn/scripts/knn/imagenet-100/knn.sh +++ /dev/null @@ -1,11 +0,0 @@ -python3 main_knn.py \ - --dataset imagenet100 \ - --train_data_path ./datasets/imagenet-100/train \ - --val_data_path ./datasets/imagenet-100/val \ - --batch_size 16 \ - --num_workers 10 \ - --pretrained_checkpoint_dir $1 \ - --k 1 2 5 10 20 50 100 200 \ - --temperature 0.01 0.02 0.05 0.07 0.1 0.2 0.5 1 \ - --feature_type backbone projector \ - --distance_function euclidean cosine diff --git a/solo-learn/scripts/linear/imagenet-100/barlow.yaml b/solo-learn/scripts/linear/cifar10/barlow.yaml similarity index 58% rename from solo-learn/scripts/linear/imagenet-100/barlow.yaml rename to solo-learn/scripts/linear/cifar10/barlow.yaml index 534859b..6a199d9 100644 --- a/solo-learn/scripts/linear/imagenet-100/barlow.yaml +++ b/solo-learn/scripts/linear/cifar10/barlow.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "barlow_twins-imagenet100-linear" -pretrained_feature_extractor: None +name: "barlow-cifar10-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" + name: "resnet50" pretrain_method: "barlow_twins" data: - dataset: imagenet100 - train_path: "/home/CORP/vg.turrisi/Documents/datasets/imagenet-100/train" - val_path: "/home/CORP/vg.turrisi/Documents/datasets/imagenet-100/val" + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH format: "image_folder" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 + name: "lars" + batch_size: 512 lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/cifar10/barlow_diff.yaml b/solo-learn/scripts/linear/cifar10/barlow_diff.yaml new file mode 100644 index 0000000..d66c394 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/barlow_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-cifar10-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/barlow_icgan.yaml b/solo-learn/scripts/linear/cifar10/barlow_icgan.yaml new file mode 100644 index 0000000..849f7ab --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/barlow_icgan.yaml @@ -0,0 +1,45 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-cifar10-icgan-linear" +pretrained_feature_extractor: MODEL_PATH + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/ressl.yaml b/solo-learn/scripts/linear/cifar10/byol.yaml similarity index 54% rename from solo-learn/scripts/linear/imagenet-100/ressl.yaml rename to solo-learn/scripts/linear/cifar10/byol.yaml index e8e87d8..09ba193 100644 --- a/solo-learn/scripts/linear/imagenet-100/ressl.yaml +++ b/solo-learn/scripts/linear/cifar10/byol.yaml @@ -10,35 +10,35 @@ hydra: run: dir: . -name: "ressl-imagenet100-linear" -pretrained_feature_extractor: None -backbone: - name: "resnet18" -pretrain_method: "ressl" +name: "byol-cifar10-linear" +pretrained_feature_extractor: MODEL_PATH + name: "resnet50" +pretrain_method: "byol" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 3.0 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/cifar10/byol_diff.yaml b/solo-learn/scripts/linear/cifar10/byol_diff.yaml new file mode 100644 index 0000000..93b7be5 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/byol_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-cifar10-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/byol_icgan.yaml b/solo-learn/scripts/linear/cifar10/byol_icgan.yaml new file mode 100644 index 0000000..659a100 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/byol_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-cifar10-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/moco.yaml b/solo-learn/scripts/linear/cifar10/moco.yaml new file mode 100644 index 0000000..a837698 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/moco.yaml @@ -0,0 +1,45 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-cifar10-linear" +pretrained_feature_extractor: MODEL_PATH + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/moco_diff.yaml b/solo-learn/scripts/linear/cifar10/moco_diff.yaml new file mode 100644 index 0000000..cdb32f9 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/moco_diff.yaml @@ -0,0 +1,45 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-cifar10-diff-linear" +pretrained_feature_extractor: MODEL_PATH + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/moco_icgan.yaml b/solo-learn/scripts/linear/cifar10/moco_icgan.yaml new file mode 100644 index 0000000..55363be --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/moco_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-cifar10-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/simclr.yaml b/solo-learn/scripts/linear/cifar10/simclr.yaml new file mode 100644 index 0000000..03aa9d8 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/simclr.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-cifar10-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/simclr_diff.yaml b/solo-learn/scripts/linear/cifar10/simclr_diff.yaml new file mode 100644 index 0000000..12b8766 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/simclr_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-cifar10-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/simclr_icgan.yaml b/solo-learn/scripts/linear/cifar10/simclr_icgan.yaml new file mode 100644 index 0000000..99ea91f --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/simclr_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-cifar10-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/simsiam.yaml b/solo-learn/scripts/linear/cifar10/simsiam.yaml new file mode 100644 index 0000000..0f01f08 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/simsiam.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-cifar10-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/simsiam_diff.yaml b/solo-learn/scripts/linear/cifar10/simsiam_diff.yaml new file mode 100644 index 0000000..0b98d90 --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/simsiam_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-cifar10-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar10/simsiam_icgan.yaml b/solo-learn/scripts/linear/cifar10/simsiam_icgan.yaml new file mode 100644 index 0000000..897c49d --- /dev/null +++ b/solo-learn/scripts/linear/cifar10/simsiam_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-cifar10-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: cifar10 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/barlow.yaml b/solo-learn/scripts/linear/cifar100/barlow.yaml new file mode 100644 index 0000000..34b5e62 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/barlow.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-cifar100-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/barlow_diff.yaml b/solo-learn/scripts/linear/cifar100/barlow_diff.yaml new file mode 100644 index 0000000..3d1f4a3 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/barlow_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-cifar100-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/barlow_icgan.yaml b/solo-learn/scripts/linear/cifar100/barlow_icgan.yaml new file mode 100644 index 0000000..41bb6f0 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/barlow_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-cifar100-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/byol.yaml b/solo-learn/scripts/linear/cifar100/byol.yaml new file mode 100644 index 0000000..fe94ae8 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/byol.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-cifar100-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/byol_diff.yaml b/solo-learn/scripts/linear/cifar100/byol_diff.yaml new file mode 100644 index 0000000..1a86a64 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/byol_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-cifar100-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/byol_icgan.yaml b/solo-learn/scripts/linear/cifar100/byol_icgan.yaml new file mode 100644 index 0000000..7ee0604 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/byol_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-cifar100-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/moco.yaml b/solo-learn/scripts/linear/cifar100/moco.yaml new file mode 100644 index 0000000..29dd632 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/moco.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-cifar100-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/moco_diff.yaml b/solo-learn/scripts/linear/cifar100/moco_diff.yaml new file mode 100644 index 0000000..867beba --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/moco_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-cifar100-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/moco_icgan.yaml b/solo-learn/scripts/linear/cifar100/moco_icgan.yaml new file mode 100644 index 0000000..e1f550e --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/moco_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-cifar100-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/simclr.yaml b/solo-learn/scripts/linear/cifar100/simclr.yaml new file mode 100644 index 0000000..9bad0b4 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/simclr.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-cifar100-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/simclr_diff.yaml b/solo-learn/scripts/linear/cifar100/simclr_diff.yaml new file mode 100644 index 0000000..1620412 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/simclr_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-cifar100-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/simclr_icgan.yaml b/solo-learn/scripts/linear/cifar100/simclr_icgan.yaml new file mode 100644 index 0000000..74ca30a --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/simclr_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-cifar100-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/simsiam.yaml b/solo-learn/scripts/linear/cifar100/simsiam.yaml new file mode 100644 index 0000000..ce86e62 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/simsiam.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-cifar100-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/simsiam_diff.yaml b/solo-learn/scripts/linear/cifar100/simsiam_diff.yaml new file mode 100644 index 0000000..3de77a2 --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/simsiam_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-cifar100-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/cifar100/simsiam_icgan.yaml b/solo-learn/scripts/linear/cifar100/simsiam_icgan.yaml new file mode 100644 index 0000000..0995d3a --- /dev/null +++ b/solo-learn/scripts/linear/cifar100/simsiam_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-cifar100-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: cifar100 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/barlow.yaml b/solo-learn/scripts/linear/food/barlow.yaml new file mode 100644 index 0000000..d2e575f --- /dev/null +++ b/solo-learn/scripts/linear/food/barlow.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-food101-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/barlow_diff.yaml b/solo-learn/scripts/linear/food/barlow_diff.yaml new file mode 100644 index 0000000..2200847 --- /dev/null +++ b/solo-learn/scripts/linear/food/barlow_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-food101-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/barlow_icgan.yaml b/solo-learn/scripts/linear/food/barlow_icgan.yaml new file mode 100644 index 0000000..c6d02e6 --- /dev/null +++ b/solo-learn/scripts/linear/food/barlow_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-food101-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/byol.yaml b/solo-learn/scripts/linear/food/byol.yaml new file mode 100644 index 0000000..1543e5c --- /dev/null +++ b/solo-learn/scripts/linear/food/byol.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-food101-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/byol_diff.yaml b/solo-learn/scripts/linear/food/byol_diff.yaml new file mode 100644 index 0000000..614809c --- /dev/null +++ b/solo-learn/scripts/linear/food/byol_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-food101-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/byol_icgan.yaml b/solo-learn/scripts/linear/food/byol_icgan.yaml new file mode 100644 index 0000000..ec08894 --- /dev/null +++ b/solo-learn/scripts/linear/food/byol_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-food101-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/moco.yaml b/solo-learn/scripts/linear/food/moco.yaml new file mode 100644 index 0000000..3fbc704 --- /dev/null +++ b/solo-learn/scripts/linear/food/moco.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-food101-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/moco_diff.yaml b/solo-learn/scripts/linear/food/moco_diff.yaml new file mode 100644 index 0000000..dadd228 --- /dev/null +++ b/solo-learn/scripts/linear/food/moco_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-food101-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/moco_icgan.yaml b/solo-learn/scripts/linear/food/moco_icgan.yaml new file mode 100644 index 0000000..77f8dee --- /dev/null +++ b/solo-learn/scripts/linear/food/moco_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-food101-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/simclr.yaml b/solo-learn/scripts/linear/food/simclr.yaml new file mode 100644 index 0000000..c66bab0 --- /dev/null +++ b/solo-learn/scripts/linear/food/simclr.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-food101-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/simclr_diff.yaml b/solo-learn/scripts/linear/food/simclr_diff.yaml new file mode 100644 index 0000000..d6528b8 --- /dev/null +++ b/solo-learn/scripts/linear/food/simclr_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-food101-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/simclr_icgan.yaml b/solo-learn/scripts/linear/food/simclr_icgan.yaml new file mode 100644 index 0000000..3550cd7 --- /dev/null +++ b/solo-learn/scripts/linear/food/simclr_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-food101-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/simsiam.yaml b/solo-learn/scripts/linear/food/simsiam.yaml new file mode 100644 index 0000000..55769a7 --- /dev/null +++ b/solo-learn/scripts/linear/food/simsiam.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-food101-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/simsiam_diff.yaml b/solo-learn/scripts/linear/food/simsiam_diff.yaml new file mode 100644 index 0000000..a224375 --- /dev/null +++ b/solo-learn/scripts/linear/food/simsiam_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-food101-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/food/simsiam_icgan.yaml b/solo-learn/scripts/linear/food/simsiam_icgan.yaml new file mode 100644 index 0000000..d4187ed --- /dev/null +++ b/solo-learn/scripts/linear/food/simsiam_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-food101-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: food101 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "image_folder" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/deepclusterv2.yaml b/solo-learn/scripts/linear/imagenet-100/deepclusterv2.yaml deleted file mode 100644 index 4d40619..0000000 --- a/solo-learn/scripts/linear/imagenet-100/deepclusterv2.yaml +++ /dev/null @@ -1,45 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "deepclusterv2-imagenet100-linear" -pretrained_feature_extractor: None -backbone: - name: "resnet18" -pretrain_method: "deepclusterv2" -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 0.15 - weight_decay: 0 -scheduler: - name: "step" - lr_decay_steps: [60, 80] -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 100 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/mocov3_vit.yaml b/solo-learn/scripts/linear/imagenet-100/mocov3_vit.yaml deleted file mode 100644 index 92a298e..0000000 --- a/solo-learn/scripts/linear/imagenet-100/mocov3_vit.yaml +++ /dev/null @@ -1,45 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mocov3-vit-imagenet100-linear" -pretrained_feature_extractor: None -backbone: - name: "vit_small" -pretrain_method: "mocov3" -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 0.3 - weight_decay: 0 -scheduler: - name: "step" - lr_decay_steps: [60, 80] -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 100 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/vibcreg.yaml b/solo-learn/scripts/linear/imagenet-100/vibcreg.yaml deleted file mode 100644 index d4ad39f..0000000 --- a/solo-learn/scripts/linear/imagenet-100/vibcreg.yaml +++ /dev/null @@ -1,45 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "vibcreg-imagenet100-linear" -pretrained_feature_extractor: None -backbone: - name: "resnet18" -pretrain_method: "vibcreg" -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 0.3 - weight_decay: 0 -scheduler: - name: "step" - lr_decay_steps: [60, 80] -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 100 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/vicreg.yaml b/solo-learn/scripts/linear/imagenet-100/vicreg.yaml deleted file mode 100644 index 0d0150b..0000000 --- a/solo-learn/scripts/linear/imagenet-100/vicreg.yaml +++ /dev/null @@ -1,45 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "vicreg-imagenet100-linear" -pretrained_feature_extractor: None -backbone: - name: "resnet18" -pretrain_method: "vicreg" -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 0.3 - weight_decay: 0 -scheduler: - name: "step" - lr_decay_steps: [60, 80] -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 100 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/wandb/mhug.yaml b/solo-learn/scripts/linear/imagenet-100/wandb/mhug.yaml deleted file mode 100644 index c842e44..0000000 --- a/solo-learn/scripts/linear/imagenet-100/wandb/mhug.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: unitn-mhug -project: "gen-ssl" diff --git a/solo-learn/scripts/linear/imagenet-100/wandb/private.yaml b/solo-learn/scripts/linear/imagenet-100/wandb/private.yaml deleted file mode 100644 index ad4e200..0000000 --- a/solo-learn/scripts/linear/imagenet-100/wandb/private.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: None -project: "gen-ssl" diff --git a/solo-learn/scripts/linear/imagenet/barlow.yaml b/solo-learn/scripts/linear/imagenet/barlow.yaml index 61d32ab..e7002d0 100644 --- a/solo-learn/scripts/linear/imagenet/barlow.yaml +++ b/solo-learn/scripts/linear/imagenet/barlow.yaml @@ -11,34 +11,35 @@ hydra: dir: . name: "barlow-imagenet-linear" -pretrained_feature_extractor: None +pretrained_feature_extractor: MODEL_PATH backbone: name: "resnet50" -pretrain_method: "barlow" +pretrain_method: "barlow_twins" data: dataset: imagenet - train_path: "./datasets/imagenet/train" - val_path: "./datasets/imagenet/val" + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 0.3 - weight_decay: 1e-5 + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/imagenet-100/dino.yaml b/solo-learn/scripts/linear/imagenet/barlow_diff.yaml similarity index 57% rename from solo-learn/scripts/linear/imagenet-100/dino.yaml rename to solo-learn/scripts/linear/imagenet/barlow_diff.yaml index edacd28..e7002d0 100644 --- a/solo-learn/scripts/linear/imagenet-100/dino.yaml +++ b/solo-learn/scripts/linear/imagenet/barlow_diff.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "dino-imagenet100-linear" -pretrained_feature_extractor: None +name: "barlow-imagenet-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" -pretrain_method: "dino" + name: "resnet50" +pretrain_method: "barlow_twins" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 0.3 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/imagenet-100/swav.yaml b/solo-learn/scripts/linear/imagenet/barlow_icgan.yaml similarity index 57% rename from solo-learn/scripts/linear/imagenet-100/swav.yaml rename to solo-learn/scripts/linear/imagenet/barlow_icgan.yaml index f0155b5..9eb768f 100644 --- a/solo-learn/scripts/linear/imagenet-100/swav.yaml +++ b/solo-learn/scripts/linear/imagenet/barlow_icgan.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "swav-imagenet100-linear" -pretrained_feature_extractor: None +name: "barlow-imagenet-icgan-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" -pretrain_method: "swav" + name: "resnet50" +pretrain_method: "barlow_twins" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 0.15 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/imagenet/byol.yaml b/solo-learn/scripts/linear/imagenet/byol.yaml index 12aef32..f8de73c 100644 --- a/solo-learn/scripts/linear/imagenet/byol.yaml +++ b/solo-learn/scripts/linear/imagenet/byol.yaml @@ -11,34 +11,35 @@ hydra: dir: . name: "byol-imagenet-linear" -pretrained_feature_extractor: None +pretrained_feature_extractor: MODEL_PATH backbone: name: "resnet50" pretrain_method: "byol" data: dataset: imagenet - train_path: "./datasets/imagenet/train" - val_path: "./datasets/imagenet/val" + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 + name: "lars" + batch_size: 512 lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/imagenet-100/nnclr.yaml b/solo-learn/scripts/linear/imagenet/byol_diff.yaml similarity index 57% rename from solo-learn/scripts/linear/imagenet-100/nnclr.yaml rename to solo-learn/scripts/linear/imagenet/byol_diff.yaml index ac197f9..d25e950 100644 --- a/solo-learn/scripts/linear/imagenet-100/nnclr.yaml +++ b/solo-learn/scripts/linear/imagenet/byol_diff.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "nnclr-imagenet100-linear" -pretrained_feature_extractor: None +name: "byol-imagenet-diff-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" -pretrain_method: "nnclr" + name: "resnet50" +pretrain_method: "byol" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 0.3 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/imagenet/byol_icgan.yaml b/solo-learn/scripts/linear/imagenet/byol_icgan.yaml new file mode 100644 index 0000000..ef0b980 --- /dev/null +++ b/solo-learn/scripts/linear/imagenet/byol_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-imagenet-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet/moco.yaml b/solo-learn/scripts/linear/imagenet/moco.yaml new file mode 100644 index 0000000..5e81863 --- /dev/null +++ b/solo-learn/scripts/linear/imagenet/moco.yaml @@ -0,0 +1,45 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-imagenet-linear" +pretrained_feature_extractor: MODEL_PATH + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/mocov2plus.yaml b/solo-learn/scripts/linear/imagenet/moco_diff.yaml similarity index 60% rename from solo-learn/scripts/linear/imagenet-100/mocov2plus.yaml rename to solo-learn/scripts/linear/imagenet/moco_diff.yaml index 55d15a0..72aaa80 100644 --- a/solo-learn/scripts/linear/imagenet-100/mocov2plus.yaml +++ b/solo-learn/scripts/linear/imagenet/moco_diff.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "mocov2plus-imagenet100-linear" -pretrained_feature_extractor: None +name: "moco-imagenet-diff-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" + name: "resnet50" pretrain_method: "mocov2plus" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 3.0 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/imagenet/moco_icgan.yaml b/solo-learn/scripts/linear/imagenet/moco_icgan.yaml new file mode 100644 index 0000000..fb15d6b --- /dev/null +++ b/solo-learn/scripts/linear/imagenet/moco_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-imagenet-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet/mocov2plus.yaml b/solo-learn/scripts/linear/imagenet/mocov2plus.yaml deleted file mode 100644 index 5840a77..0000000 --- a/solo-learn/scripts/linear/imagenet/mocov2plus.yaml +++ /dev/null @@ -1,45 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mocov2plus-imagenet-linear" -pretrained_feature_extractor: "trained_models/mocov2plus/gjf2upj4/mocov2plus-imagenet-gjf2upj4-ep=99.ckpt" -backbone: - name: "resnet50" -pretrain_method: "mocov2plus" -data: - dataset: imagenet - train_path: "./datasets/imagenet/train" - val_path: "./datasets/imagenet/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 3.0 - weight_decay: 0 -scheduler: - name: "step" - lr_decay_steps: [60, 80] -checkpoint: - enabled: True - dir: "/projects/imagenet_synthetic/model_checkpoints/solo-learn/solo_trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 100 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/simclr.yaml b/solo-learn/scripts/linear/imagenet/simclr.yaml similarity index 60% rename from solo-learn/scripts/linear/imagenet-100/simclr.yaml rename to solo-learn/scripts/linear/imagenet/simclr.yaml index 04e312f..6e3bc09 100644 --- a/solo-learn/scripts/linear/imagenet-100/simclr.yaml +++ b/solo-learn/scripts/linear/imagenet/simclr.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "simclr-imagenet100-linear" -pretrained_feature_extractor: None +name: "simclr-imagenet-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" + name: "resnet50" pretrain_method: "simclr" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 1.0 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/imagenet/simclr_diff.yaml b/solo-learn/scripts/linear/imagenet/simclr_diff.yaml new file mode 100644 index 0000000..449edd1 --- /dev/null +++ b/solo-learn/scripts/linear/imagenet/simclr_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-imagenet-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet/simclr_icgan.yaml b/solo-learn/scripts/linear/imagenet/simclr_icgan.yaml new file mode 100644 index 0000000..dc67597 --- /dev/null +++ b/solo-learn/scripts/linear/imagenet/simclr_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-imagenet-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet/simsiam.yaml b/solo-learn/scripts/linear/imagenet/simsiam.yaml index 44206bf..79eb529 100644 --- a/solo-learn/scripts/linear/imagenet/simsiam.yaml +++ b/solo-learn/scripts/linear/imagenet/simsiam.yaml @@ -10,31 +10,32 @@ hydra: run: dir: . -name: "simsiam-linear" -pretrained_feature_extractor: "/projects/imagenet_synthetic/model_checkpoints/solo-learn/trained_models/simsiam/5/simsiam-imagenet-5-ep=99.ckpt" +name: "simsiam-imagenet-linear" +pretrained_feature_extractor: MODEL_PATH backbone: name: "resnet50" pretrain_method: "simsiam" data: dataset: imagenet - train_path: "/datasets/imagenet/train" - val_path: "/datasets/imagenet/val" + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 30.0 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "/projects/imagenet_synthetic/model_checkpoints/solo-learn/solo_trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: - enabled: False + enabled: True # overwrite PL stuff max_epochs: 100 diff --git a/solo-learn/scripts/linear/imagenet/simsiam_diff.yaml b/solo-learn/scripts/linear/imagenet/simsiam_diff.yaml new file mode 100644 index 0000000..a3b7d56 --- /dev/null +++ b/solo-learn/scripts/linear/imagenet/simsiam_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-imagenet-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet/simsiam_icgan.yaml b/solo-learn/scripts/linear/imagenet/simsiam_icgan.yaml new file mode 100644 index 0000000..e060243 --- /dev/null +++ b/solo-learn/scripts/linear/imagenet/simsiam_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-imagenet-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/barlow.yaml b/solo-learn/scripts/linear/inaturalist/barlow.yaml new file mode 100644 index 0000000..38fb46d --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/barlow.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-inaturalist-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/barlow_diff.yaml b/solo-learn/scripts/linear/inaturalist/barlow_diff.yaml new file mode 100644 index 0000000..38fb46d --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/barlow_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-inaturalist-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/mocov3.yaml b/solo-learn/scripts/linear/inaturalist/barlow_icgan.yaml similarity index 57% rename from solo-learn/scripts/linear/imagenet-100/mocov3.yaml rename to solo-learn/scripts/linear/inaturalist/barlow_icgan.yaml index 30beaf1..8f0253d 100644 --- a/solo-learn/scripts/linear/imagenet-100/mocov3.yaml +++ b/solo-learn/scripts/linear/inaturalist/barlow_icgan.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "mocov3-imagenet100-linear" -pretrained_feature_extractor: None +name: "barlow-inaturalist-icgan-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" -pretrain_method: "mocov3" + name: "resnet50" +pretrain_method: "barlow_twins" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 0.3 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/imagenet-100/byol.yaml b/solo-learn/scripts/linear/inaturalist/byol.yaml similarity index 60% rename from solo-learn/scripts/linear/imagenet-100/byol.yaml rename to solo-learn/scripts/linear/inaturalist/byol.yaml index e167722..dcca72a 100644 --- a/solo-learn/scripts/linear/imagenet-100/byol.yaml +++ b/solo-learn/scripts/linear/inaturalist/byol.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "byol-imagenet100-linear" -pretrained_feature_extractor: None +name: "byol-inaturalist-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" + name: "resnet50" pretrain_method: "byol" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 0.3 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/inaturalist/byol_diff.yaml b/solo-learn/scripts/linear/inaturalist/byol_diff.yaml new file mode 100644 index 0000000..1fb3c01 --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/byol_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-inaturalist-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/byol_icgan.yaml b/solo-learn/scripts/linear/inaturalist/byol_icgan.yaml new file mode 100644 index 0000000..53f169b --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/byol_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-inaturalist-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/moco.yaml b/solo-learn/scripts/linear/inaturalist/moco.yaml new file mode 100644 index 0000000..97a3239 --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/moco.yaml @@ -0,0 +1,45 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-inaturalist-linear" +pretrained_feature_extractor: MODEL_PATH + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/moco_diff.yaml b/solo-learn/scripts/linear/inaturalist/moco_diff.yaml new file mode 100644 index 0000000..936afbf --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/moco_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-inaturalist-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/moco_icgan.yaml b/solo-learn/scripts/linear/inaturalist/moco_icgan.yaml new file mode 100644 index 0000000..5e139c3 --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/moco_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-inaturalist-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/simclr.yaml b/solo-learn/scripts/linear/inaturalist/simclr.yaml new file mode 100644 index 0000000..8f6862d --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/simclr.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-inaturalist-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/simclr_diff.yaml b/solo-learn/scripts/linear/inaturalist/simclr_diff.yaml new file mode 100644 index 0000000..7abb55b --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/simclr_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-inaturalist-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/simclr_icgan.yaml b/solo-learn/scripts/linear/inaturalist/simclr_icgan.yaml new file mode 100644 index 0000000..9c72052 --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/simclr_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-inaturalist-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/imagenet-100/simsiam.yaml b/solo-learn/scripts/linear/inaturalist/simsiam.yaml similarity index 60% rename from solo-learn/scripts/linear/imagenet-100/simsiam.yaml rename to solo-learn/scripts/linear/inaturalist/simsiam.yaml index b7d9dda..3ebd8db 100644 --- a/solo-learn/scripts/linear/imagenet-100/simsiam.yaml +++ b/solo-learn/scripts/linear/inaturalist/simsiam.yaml @@ -10,35 +10,36 @@ hydra: run: dir: . -name: "simsiam-imagenet100-linear" -pretrained_feature_extractor: None +name: "simsiam-inaturalist-linear" +pretrained_feature_extractor: MODEL_PATH backbone: - name: "resnet18" + name: "resnet50" pretrain_method: "simsiam" data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: - name: "sgd" - batch_size: 256 - lr: 30.0 + name: "lars" + batch_size: 512 + lr: 0.1 weight_decay: 0 scheduler: - name: "step" - lr_decay_steps: [60, 80] + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff max_epochs: 100 -devices: [0] +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/linear/inaturalist/simsiam_diff.yaml b/solo-learn/scripts/linear/inaturalist/simsiam_diff.yaml new file mode 100644 index 0000000..121f20e --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/simsiam_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-inaturalist-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/inaturalist/simsiam_icgan.yaml b/solo-learn/scripts/linear/inaturalist/simsiam_icgan.yaml new file mode 100644 index 0000000..7ede227 --- /dev/null +++ b/solo-learn/scripts/linear/inaturalist/simsiam_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-inaturalist-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: inaturalist + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/barlow.yaml b/solo-learn/scripts/linear/places/barlow.yaml new file mode 100644 index 0000000..ccba6d6 --- /dev/null +++ b/solo-learn/scripts/linear/places/barlow.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-places-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/barlow_diff.yaml b/solo-learn/scripts/linear/places/barlow_diff.yaml new file mode 100644 index 0000000..3b819f0 --- /dev/null +++ b/solo-learn/scripts/linear/places/barlow_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-places-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/barlow_icgan.yaml b/solo-learn/scripts/linear/places/barlow_icgan.yaml new file mode 100644 index 0000000..ae67213 --- /dev/null +++ b/solo-learn/scripts/linear/places/barlow_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "barlow-places-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "barlow_twins" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/byol.yaml b/solo-learn/scripts/linear/places/byol.yaml new file mode 100644 index 0000000..89faf1d --- /dev/null +++ b/solo-learn/scripts/linear/places/byol.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-places-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/byol_diff.yaml b/solo-learn/scripts/linear/places/byol_diff.yaml new file mode 100644 index 0000000..0af3f6a --- /dev/null +++ b/solo-learn/scripts/linear/places/byol_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-places-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/byol_icgan.yaml b/solo-learn/scripts/linear/places/byol_icgan.yaml new file mode 100644 index 0000000..3f3829f --- /dev/null +++ b/solo-learn/scripts/linear/places/byol_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "byol-places-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "byol" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/moco.yaml b/solo-learn/scripts/linear/places/moco.yaml new file mode 100644 index 0000000..e02643e --- /dev/null +++ b/solo-learn/scripts/linear/places/moco.yaml @@ -0,0 +1,45 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-places-linear" +pretrained_feature_extractor: MODEL_PATH + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/moco_diff.yaml b/solo-learn/scripts/linear/places/moco_diff.yaml new file mode 100644 index 0000000..a2e1596 --- /dev/null +++ b/solo-learn/scripts/linear/places/moco_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-places-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/moco_icgan.yaml b/solo-learn/scripts/linear/places/moco_icgan.yaml new file mode 100644 index 0000000..675e59f --- /dev/null +++ b/solo-learn/scripts/linear/places/moco_icgan.yaml @@ -0,0 +1,45 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "moco-places-icgan-linear" +pretrained_feature_extractor: MODEL_PATH + name: "resnet50" +pretrain_method: "mocov2plus" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/simclr.yaml b/solo-learn/scripts/linear/places/simclr.yaml new file mode 100644 index 0000000..1c010d4 --- /dev/null +++ b/solo-learn/scripts/linear/places/simclr.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-places-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/simclr_diff.yaml b/solo-learn/scripts/linear/places/simclr_diff.yaml new file mode 100644 index 0000000..0352a35 --- /dev/null +++ b/solo-learn/scripts/linear/places/simclr_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-places-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/simclr_icgan.yaml b/solo-learn/scripts/linear/places/simclr_icgan.yaml new file mode 100644 index 0000000..939dd73 --- /dev/null +++ b/solo-learn/scripts/linear/places/simclr_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-places-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simclr" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/simsiam.yaml b/solo-learn/scripts/linear/places/simsiam.yaml new file mode 100644 index 0000000..1cf2301 --- /dev/null +++ b/solo-learn/scripts/linear/places/simsiam.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-places-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/simsiam_diff.yaml b/solo-learn/scripts/linear/places/simsiam_diff.yaml new file mode 100644 index 0000000..789d03a --- /dev/null +++ b/solo-learn/scripts/linear/places/simsiam_diff.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-places-diff-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/linear/places/simsiam_icgan.yaml b/solo-learn/scripts/linear/places/simsiam_icgan.yaml new file mode 100644 index 0000000..ad79c48 --- /dev/null +++ b/solo-learn/scripts/linear/places/simsiam_icgan.yaml @@ -0,0 +1,46 @@ +defaults: + - _self_ + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simsiam-places-icgan-linear" +pretrained_feature_extractor: MODEL_PATH +backbone: + name: "resnet50" +pretrain_method: "simsiam" +data: + dataset: places365 + train_path: TRAIN_PATH + val_path: VAL_PATH + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 512 + lr: 0.1 + weight_decay: 0 +scheduler: + name: "warmup_cosine" + warmup_epochs: 0 + scheduler_interval: "epoch" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 45 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16 diff --git a/solo-learn/scripts/pretrain/cifar-multicrop/augmentations/swav.yaml b/solo-learn/scripts/pretrain/cifar-multicrop/augmentations/swav.yaml deleted file mode 100644 index 95baaf3..0000000 --- a/solo-learn/scripts/pretrain/cifar-multicrop/augmentations/swav.yaml +++ /dev/null @@ -1,68 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.6 - contrast: 0.6 - saturation: 0.6 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 1 - -- rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.6 - contrast: 0.6 - saturation: 0.6 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.2 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 1 - -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 0.5 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.1 - solarization: - prob: 0.1 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 18 - num_crops: 6 diff --git a/solo-learn/scripts/pretrain/cifar-multicrop/swav.yaml b/solo-learn/scripts/pretrain/cifar-multicrop/swav.yaml deleted file mode 100644 index c36b766..0000000 --- a/solo-learn/scripts/pretrain/cifar-multicrop/swav.yaml +++ /dev/null @@ -1,58 +0,0 @@ -defaults: - - _self_ - - augmentations: swav.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "swav-cifar10-multicrop" # change here for cifar100 -method: "swav" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - queue_size: 3840 - proj_output_dim: 128 - num_prototypes: 3000 - epoch_queue_starts: 50 - freeze_prototypes_epochs: 2 - temperature: 0.1 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "datasets/imagenet100/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.6 - classifier_lr: 0.1 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" - min_lr: 0.0006 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar-multicrop/wandb/mhug.yaml b/solo-learn/scripts/pretrain/cifar-multicrop/wandb/mhug.yaml deleted file mode 100644 index c842e44..0000000 --- a/solo-learn/scripts/pretrain/cifar-multicrop/wandb/mhug.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: unitn-mhug -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/cifar-multicrop/wandb/private.yaml b/solo-learn/scripts/pretrain/cifar-multicrop/wandb/private.yaml deleted file mode 100644 index ad4e200..0000000 --- a/solo-learn/scripts/pretrain/cifar-multicrop/wandb/private.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: None -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/cifar/all4one.yaml b/solo-learn/scripts/pretrain/cifar/all4one.yaml deleted file mode 100644 index 7db7ea7..0000000 --- a/solo-learn/scripts/pretrain/cifar/all4one.yaml +++ /dev/null @@ -1,58 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "All4One-cifar100" # change here for cifar10 -method: "all4one" -backbone: - name: "resnet18" -method_kwargs: - temperature: 0.2 - proj_hidden_dim: 2048 - pred_hidden_dim: 4096 - proj_output_dim: 256 - queue_size: 98304 -momentum: - base_tau: 0.99 - final_tau: 1.0 -data: - dataset: cifar100 # change here for cifar10 - train_path: "./datasets/" - val_path: "./datasets/" - format: "image_folder" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 256 - lr: 1.0 - classifier_lr: 0.1 - weight_decay: 1e-5 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: False - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/augmentations/asymmetric.yaml b/solo-learn/scripts/pretrain/cifar/augmentations/asymmetric.yaml deleted file mode 100644 index 8eb4ebc..0000000 --- a/solo-learn/scripts/pretrain/cifar/augmentations/asymmetric.yaml +++ /dev/null @@ -1,45 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 1 - -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.2 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 1 diff --git a/solo-learn/scripts/pretrain/cifar/augmentations/reconstruction.yaml b/solo-learn/scripts/pretrain/cifar/augmentations/reconstruction.yaml deleted file mode 100644 index 56e7549..0000000 --- a/solo-learn/scripts/pretrain/cifar/augmentations/reconstruction.yaml +++ /dev/null @@ -1,18 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.0 - grayscale: - prob: 0.0 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 1 diff --git a/solo-learn/scripts/pretrain/cifar/augmentations/ressl.yaml b/solo-learn/scripts/pretrain/cifar/augmentations/ressl.yaml deleted file mode 100644 index 7f6dc2c..0000000 --- a/solo-learn/scripts/pretrain/cifar/augmentations/ressl.yaml +++ /dev/null @@ -1,41 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 1 - -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.0 - grayscale: - prob: 0.0 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 1 diff --git a/solo-learn/scripts/pretrain/cifar/augmentations/symmetric.yaml b/solo-learn/scripts/pretrain/cifar/augmentations/symmetric.yaml deleted file mode 100644 index 24078d9..0000000 --- a/solo-learn/scripts/pretrain/cifar/augmentations/symmetric.yaml +++ /dev/null @@ -1,22 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 2 diff --git a/solo-learn/scripts/pretrain/cifar/augmentations/symmetric_weak.yaml b/solo-learn/scripts/pretrain/cifar/augmentations/symmetric_weak.yaml deleted file mode 100644 index 8ce8159..0000000 --- a/solo-learn/scripts/pretrain/cifar/augmentations/symmetric_weak.yaml +++ /dev/null @@ -1,22 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.4 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 2 diff --git a/solo-learn/scripts/pretrain/cifar/barlow.yaml b/solo-learn/scripts/pretrain/cifar/barlow.yaml deleted file mode 100644 index 728f14b..0000000 --- a/solo-learn/scripts/pretrain/cifar/barlow.yaml +++ /dev/null @@ -1,53 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "barlow_twins-cifar10" # change here for cifar100 -method: "barlow_twins" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 2048 - scale_loss: 0.1 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-4 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/deepclusterv2.yaml b/solo-learn/scripts/pretrain/cifar/deepclusterv2.yaml deleted file mode 100644 index f884785..0000000 --- a/solo-learn/scripts/pretrain/cifar/deepclusterv2.yaml +++ /dev/null @@ -1,56 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "deepclusterv2-cifar10" # change here for cifar100 -method: "deepclusterv2" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 128 - num_prototypes: [3000, 3000, 3000] -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.6 - classifier_lr: 0.1 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" - min_lr: 0.0006 - warmup_start_lr: 0.0 - warmup_epochs: 11 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/dino.yaml b/solo-learn/scripts/pretrain/cifar/dino.yaml deleted file mode 100644 index 008e3ab..0000000 --- a/solo-learn/scripts/pretrain/cifar/dino.yaml +++ /dev/null @@ -1,56 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "dino-cifar10" # change here for cifar100 -method: "dino" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 256 - num_prototypes: 4096 -momentum: - base_tau: 0.9995 - final_tau: 1.0 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/mae.yaml b/solo-learn/scripts/pretrain/cifar/mae.yaml deleted file mode 100644 index 0d8f8ba..0000000 --- a/solo-learn/scripts/pretrain/cifar/mae.yaml +++ /dev/null @@ -1,56 +0,0 @@ -defaults: - - _self_ - - augmentations: reconstruction.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mae-cifar10" # change here for cifar100 -method: "mae" -backbone: - name: "vit_small" - kwargs: - patch_size: 4 - img_size: 32 -method_kwargs: - decoder_embed_dim: 512 - decoder_depth: 8 - decoder_num_heads: 16 - mask_ratio: 0.75 - norm_pix_loss: True -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" - num_workers: 4 -optimizer: - name: "adamw" - batch_size: 256 - lr: 2.0e-4 - classifier_lr: 2.0e-4 - weight_decay: 0.05 - kwargs: - betas: [0.9, 0.95] -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/nnclr.yaml b/solo-learn/scripts/pretrain/cifar/nnclr.yaml deleted file mode 100644 index 2786f36..0000000 --- a/solo-learn/scripts/pretrain/cifar/nnclr.yaml +++ /dev/null @@ -1,55 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "nnclr-cifar10" # change here for cifar100 -method: "nnclr" -backbone: - name: "resnet18" -method_kwargs: - temperature: 0.2 - proj_hidden_dim: 2048 - pred_hidden_dim: 4096 - proj_output_dim: 256 - queue_size: 65536 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.4 - classifier_lr: 0.1 - weight_decay: 1e-5 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/nnsiam.yaml b/solo-learn/scripts/pretrain/cifar/nnsiam.yaml deleted file mode 100644 index 3d611e7..0000000 --- a/solo-learn/scripts/pretrain/cifar/nnsiam.yaml +++ /dev/null @@ -1,53 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "nnsiam-cifar10" # change here for cifar100 -method: "nnsiam" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - pred_hidden_dim: 4096 - proj_output_dim: 2048 - queue_size: 65536 -momentum: - base_tau: 0.99 - final_tau: 1.0 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 0.5 - classifier_lr: 0.1 - weight_decay: 1e-5 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/ressl.yaml b/solo-learn/scripts/pretrain/cifar/ressl.yaml deleted file mode 100644 index 7272f62..0000000 --- a/solo-learn/scripts/pretrain/cifar/ressl.yaml +++ /dev/null @@ -1,56 +0,0 @@ -defaults: - - _self_ - - augmentations: ressl.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "ressl-cifar10" # change here for cifar100 -method: "ressl" -backbone: - name: "resnet18" -method_kwargs: - proj_output_dim: 256 - proj_hidden_dim: 4096 - base_tau_momentum: 0.99 - final_tau_momentum: 1.0 - momentum_classifier: - temperature_q: 0.1 - temperature_k: 0.04 -momentum: - base_tau: 0.99 - final_tau: 1.0 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 0.05 - classifier_lr: 0.1 - weight_decay: 1e-4 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/simclr.yaml b/solo-learn/scripts/pretrain/cifar/simclr.yaml deleted file mode 100644 index 0531365..0000000 --- a/solo-learn/scripts/pretrain/cifar/simclr.yaml +++ /dev/null @@ -1,53 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "simclr-cifar10" # change here for cifar100 -method: "simclr" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 256 - temperature: 0.2 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "datasets/imagenet100/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.4 - classifier_lr: 0.1 - weight_decay: 1e-4 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/simsiam.yaml b/solo-learn/scripts/pretrain/cifar/simsiam.yaml deleted file mode 100644 index dec94d4..0000000 --- a/solo-learn/scripts/pretrain/cifar/simsiam.yaml +++ /dev/null @@ -1,50 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric_weak.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "simsiam-cifar10" # change here for cifar100 -method: "simsiam" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 2048 - pred_hidden_dim: 512 - temperature: 0.2 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "datasets/imagenet100/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 0.5 - classifier_lr: 0.1 - weight_decay: 1e-5 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/supcon.yaml b/solo-learn/scripts/pretrain/cifar/supcon.yaml deleted file mode 100644 index 365317b..0000000 --- a/solo-learn/scripts/pretrain/cifar/supcon.yaml +++ /dev/null @@ -1,49 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "supcon-cifar10" # change here for cifar100 -method: "supcon" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 256 - temperature: 0.2 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "datasets/imagenet100/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 256 - lr: 0.4 - classifier_lr: 0.1 - weight_decay: 1e-5 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/swav.yaml b/solo-learn/scripts/pretrain/cifar/swav.yaml deleted file mode 100644 index 01d6c43..0000000 --- a/solo-learn/scripts/pretrain/cifar/swav.yaml +++ /dev/null @@ -1,57 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "swav-cifar10" # change here for cifar100 -method: "swav" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - queue_size: 3840 - proj_output_dim: 128 - num_prototypes: 3000 - epoch_queue_starts: 50 - freeze_prototypes_epochs: 2 - temperature: 0.1 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "datasets/imagenet100/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.6 - classifier_lr: 0.1 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/vibcreg.yaml b/solo-learn/scripts/pretrain/cifar/vibcreg.yaml deleted file mode 100644 index ebc2404..0000000 --- a/solo-learn/scripts/pretrain/cifar/vibcreg.yaml +++ /dev/null @@ -1,77 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "vibcreg-cifar10" # change here for cifar100 -method: "vibcreg" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 2048 - sim_loss_weight: 25.0 - var_loss_weight: 25.0 - cov_loss_weight: 200.0 - iternorm: True -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "datasets/imagenet100/val" - format: "image_folder" - num_workers: 4 -augmentations: - - rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.1 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 32 - num_crops: 2 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-4 - kwargs: - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/vicreg.yaml b/solo-learn/scripts/pretrain/cifar/vicreg.yaml deleted file mode 100644 index 0a8db31..0000000 --- a/solo-learn/scripts/pretrain/cifar/vicreg.yaml +++ /dev/null @@ -1,83 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "vicreg-cifar10" # change here for cifar100 -method: "vicreg" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 2048 - sim_loss_weight: 25.0 - var_loss_weight: 25.0 - cov_loss_weight: 1.0 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "datasets/imagenet100/val" - format: "image_folder" - num_workers: 4 -augmentations: - - rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - enabled: True - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - prob: 0.8 - grayscale: - enabled: True - prob: 0.2 - gaussian_blur: - enabled: False - prob: 0.0 - solarization: - enabled: True - prob: 0.1 - equalization: - enabled: False - prob: 0.0 - horizontal_flip: - enabled: True - prob: 0.5 - crop_size: 32 - num_crops: 2 -optimizer: - name: "lars" - batch_size: 256 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-4 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/cifar/wandb/mhug.yaml b/solo-learn/scripts/pretrain/cifar/wandb/mhug.yaml deleted file mode 100644 index c842e44..0000000 --- a/solo-learn/scripts/pretrain/cifar/wandb/mhug.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: unitn-mhug -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/cifar/wandb/private.yaml b/solo-learn/scripts/pretrain/cifar/wandb/private.yaml deleted file mode 100644 index ad4e200..0000000 --- a/solo-learn/scripts/pretrain/cifar/wandb/private.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: None -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/cifar/wmse.yaml b/solo-learn/scripts/pretrain/cifar/wmse.yaml deleted file mode 100644 index 7b77e45..0000000 --- a/solo-learn/scripts/pretrain/cifar/wmse.yaml +++ /dev/null @@ -1,73 +0,0 @@ -defaults: - - _self_ - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "wmse-cifar10" # change here for cifar100 -method: "wmse" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 1024 - proj_output_dim: 64 - whitening_size: 128 -data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "datasets/imagenet100/val" - format: "image_folder" - num_workers: 4 -augmentations: - - rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - enabled: True - prob: 0.5 - crop_size: 32 - num_crops: 2 -optimizer: - name: "adam" - batch_size: 256 - lr: 2e-3 - classifier_lr: 3e-3 - weight_decay: 1e-6 -scheduler: - name: "warmup_cosine" - warmup_start_lr: 0 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 1000 -devices: [0] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/custom/augmentations/asymmetric.yaml b/solo-learn/scripts/pretrain/custom/augmentations/asymmetric.yaml deleted file mode 100644 index 30d8d26..0000000 --- a/solo-learn/scripts/pretrain/custom/augmentations/asymmetric.yaml +++ /dev/null @@ -1,45 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 1.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 1 - -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.1 - solarization: - prob: 0.2 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 1 diff --git a/solo-learn/scripts/pretrain/custom/augmentations/symmetric.yaml b/solo-learn/scripts/pretrain/custom/augmentations/symmetric.yaml deleted file mode 100644 index a852a08..0000000 --- a/solo-learn/scripts/pretrain/custom/augmentations/symmetric.yaml +++ /dev/null @@ -1,22 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 2 diff --git a/solo-learn/scripts/pretrain/custom/byol.yaml b/solo-learn/scripts/pretrain/custom/byol.yaml deleted file mode 100644 index 517dcb4..0000000 --- a/solo-learn/scripts/pretrain/custom/byol.yaml +++ /dev/null @@ -1,63 +0,0 @@ - -# how to configure the augmentations -# it's also possible to copy paste here for a finer control -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "byol-custom-dataset" -method: "byol" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 4096 - proj_output_dim: 256 - pred_hidden_dim: 8192 -momentum: - base_tau: 0.99 - final_tau: 1.0 -data: - dataset: "custom" - train_path: "PATH_TO_TRAIN_DIR" - val_path: "PATH_TO_VAL_DIR" # remove this if there's no validation dir - format: "dali" # data format, supports "image_folder", "dali" or "h5" - num_workers: 4 - # set this to True if the dataset is not stored as subfolders for each class - # if no labels are provided, "h5" is not supported - # convert a custom dataset by following `scripts/utils/convert_imgfolder_to_h5.py` - no_labels: False -optimizer: - name: "lars" - batch_size: 64 - lr: 0.5 - classifier_lr: 0.1 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/custom/wandb/mhug.yaml b/solo-learn/scripts/pretrain/custom/wandb/mhug.yaml deleted file mode 100644 index c842e44..0000000 --- a/solo-learn/scripts/pretrain/custom/wandb/mhug.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: unitn-mhug -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/custom/wandb/private.yaml b/solo-learn/scripts/pretrain/custom/wandb/private.yaml deleted file mode 100644 index ad4e200..0000000 --- a/solo-learn/scripts/pretrain/custom/wandb/private.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: None -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/imagenet-100-multicrop/augmentations/asymmetric.yaml b/solo-learn/scripts/pretrain/imagenet-100-multicrop/augmentations/asymmetric.yaml deleted file mode 100644 index b5f253f..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100-multicrop/augmentations/asymmetric.yaml +++ /dev/null @@ -1,68 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 1.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 1 - -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.1 - solarization: - prob: 0.2 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 1 - -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 96 - num_crops: 6 diff --git a/solo-learn/scripts/pretrain/imagenet-100-multicrop/augmentations/symmetric.yaml b/solo-learn/scripts/pretrain/imagenet-100-multicrop/augmentations/symmetric.yaml deleted file mode 100644 index 8ca7032..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100-multicrop/augmentations/symmetric.yaml +++ /dev/null @@ -1,45 +0,0 @@ - - rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 2 - - - rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 96 - num_crops: 6 diff --git a/solo-learn/scripts/pretrain/imagenet-100-multicrop/supcon.yaml b/solo-learn/scripts/pretrain/imagenet-100-multicrop/supcon.yaml deleted file mode 100644 index 5de0a77..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100-multicrop/supcon.yaml +++ /dev/null @@ -1,49 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "supcon-multicrop-imagenet100" -method: "supcon" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 512 - temperature: 0.2 -data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 64 - lr: 0.5 - classifier_lr: 0.1 - weight_decay: 1e-5 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100-multicrop/wandb/mhug.yaml b/solo-learn/scripts/pretrain/imagenet-100-multicrop/wandb/mhug.yaml deleted file mode 100644 index c842e44..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100-multicrop/wandb/mhug.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: unitn-mhug -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/imagenet-100-multicrop/wandb/private.yaml b/solo-learn/scripts/pretrain/imagenet-100-multicrop/wandb/private.yaml deleted file mode 100644 index ad4e200..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100-multicrop/wandb/private.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: None -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/imagenet-100/all4one.yml b/solo-learn/scripts/pretrain/imagenet-100/all4one.yml deleted file mode 100644 index 8cf76c8..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/all4one.yml +++ /dev/null @@ -1,55 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "all4one-imagenet100" -method: "all4one" -backbone: - name: "resnet18" -method_kwargs: - temperature: 0.2 - proj_hidden_dim: 2048 - pred_hidden_dim: 4096 - proj_output_dim: 256 - queue_size: 98340 -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 128 - lr: 1.0 - classifier_lr: 0.1 - weight_decay: 1e-5 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/augmentations/reconstruction.yaml b/solo-learn/scripts/pretrain/imagenet-100/augmentations/reconstruction.yaml deleted file mode 100644 index 2ebd9fa..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/augmentations/reconstruction.yaml +++ /dev/null @@ -1,18 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.0 - grayscale: - prob: 0.0 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 1 diff --git a/solo-learn/scripts/pretrain/imagenet-100/augmentations/ressl.yaml b/solo-learn/scripts/pretrain/imagenet-100/augmentations/ressl.yaml deleted file mode 100644 index 328b15a..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/augmentations/ressl.yaml +++ /dev/null @@ -1,41 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 1 - -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.0 - grayscale: - prob: 0.0 - gaussian_blur: - prob: 0.0 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 1 diff --git a/solo-learn/scripts/pretrain/imagenet-100/augmentations/symmetric.yaml b/solo-learn/scripts/pretrain/imagenet-100/augmentations/symmetric.yaml deleted file mode 100644 index a852a08..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/augmentations/symmetric.yaml +++ /dev/null @@ -1,22 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 2 diff --git a/solo-learn/scripts/pretrain/imagenet-100/augmentations/symmetric_weak.yaml b/solo-learn/scripts/pretrain/imagenet-100/augmentations/symmetric_weak.yaml deleted file mode 100644 index 921d7dc..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/augmentations/symmetric_weak.yaml +++ /dev/null @@ -1,22 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.08 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.4 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 2 diff --git a/solo-learn/scripts/pretrain/imagenet-100/augmentations/vicreg.yaml b/solo-learn/scripts/pretrain/imagenet-100/augmentations/vicreg.yaml deleted file mode 100644 index 05ec827..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/augmentations/vicreg.yaml +++ /dev/null @@ -1,22 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.5 - solarization: - prob: 0.1 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 2 diff --git a/solo-learn/scripts/pretrain/imagenet-100/augmentations/wmse.yaml b/solo-learn/scripts/pretrain/imagenet-100/augmentations/wmse.yaml deleted file mode 100644 index 423e691..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/augmentations/wmse.yaml +++ /dev/null @@ -1,45 +0,0 @@ -- rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.2 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 224 - num_crops: 2 - -- rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - prob: 0.8 - brightness: 0.8 - contrast: 0.8 - saturation: 0.8 - hue: 0.2 - grayscale: - prob: 0.2 - gaussian_blur: - prob: 0.2 - solarization: - prob: 0.0 - equalization: - prob: 0.0 - horizontal_flip: - prob: 0.5 - crop_size: 96 - num_crops: 6 diff --git a/solo-learn/scripts/pretrain/imagenet-100/barlow.yaml b/solo-learn/scripts/pretrain/imagenet-100/barlow.yaml deleted file mode 100644 index ddd2da6..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/barlow.yaml +++ /dev/null @@ -1,53 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "barlow_twins-imagenet100" -method: "barlow_twins" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 2048 - scale_loss: 0.1 -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 128 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-4 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/deepclusterv2.yaml b/solo-learn/scripts/pretrain/imagenet-100/deepclusterv2.yaml deleted file mode 100644 index f6c023f..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/deepclusterv2.yaml +++ /dev/null @@ -1,58 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "deepclusterv2-imagenet100" -method: "deepclusterv2" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 128 - num_prototypes: [3000, 3000, 3000] -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 128 - lr: 0.6 - classifier_lr: 0.1 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" - min_lr: 0.0006 - warmup_start_lr: 0.0 - warmup_epochs: 11 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True -dali: - encode_indexes_into_labels: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/dino.yaml b/solo-learn/scripts/pretrain/imagenet-100/dino.yaml deleted file mode 100644 index 1129e12..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/dino.yaml +++ /dev/null @@ -1,57 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "dino-imagenet100" -method: "dino" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 256 - num_prototypes: 4096 - warmup_teacher_temperature_epochs: 50 -momentum: - base_tau: 0.9995 - final_tau: 1.0 -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 128 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/dino_vit.yaml b/solo-learn/scripts/pretrain/imagenet-100/dino_vit.yaml deleted file mode 100644 index 89ff43b..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/dino_vit.yaml +++ /dev/null @@ -1,54 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "dino-vit-imagenet100" -method: "dino" -backbone: - name: "vit_tiny" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 256 - num_prototypes: 65536 - norm_last_layer: False -momentum: - base_tau: 0.9995 - final_tau: 1.0 -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "adamw" - batch_size: 128 - lr: 0.005 - classifier_lr: 3e-3 - weight_decay: 1e-4 -scheduler: - name: "warmup_cosine" - warmup_start_lr: 0.00001 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/mae.yaml b/solo-learn/scripts/pretrain/imagenet-100/mae.yaml deleted file mode 100644 index 7366cd6..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/mae.yaml +++ /dev/null @@ -1,54 +0,0 @@ -defaults: - - _self_ - - augmentations: reconstruction.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mae-imagenet100" -method: "mae" -backbone: - name: "vit_base" -method_kwargs: - decoder_embed_dim: 512 - decoder_depth: 8 - decoder_num_heads: 16 - mask_ratio: 0.75 - norm_pix_loss: True -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "adamw" - batch_size: 128 - lr: 2.0e-4 - classifier_lr: 2.0e-4 - weight_decay: 0.05 - kwargs: - betas: [0.9, 0.95] -scheduler: - name: "warmup_cosine" - warmup_start_lr: 0.0 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/mocov3.yaml b/solo-learn/scripts/pretrain/imagenet-100/mocov3.yaml deleted file mode 100644 index df5d4de..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/mocov3.yaml +++ /dev/null @@ -1,57 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mocov3-imagenet100" -method: "mocov3" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 4096 - proj_output_dim: 256 - pred_hidden_dim: 4096 - temperature: 0.2 -momentum: - base_tau: 0.99 - final_tau: 1.0 -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 128 - lr: 0.3 - classifier_lr: 0.3 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/mocov3_vit.yaml b/solo-learn/scripts/pretrain/imagenet-100/mocov3_vit.yaml deleted file mode 100644 index af942c5..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/mocov3_vit.yaml +++ /dev/null @@ -1,53 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mocov3-vit-imagenet100" -method: "mocov3" -backbone: - name: "vit_small" -method_kwargs: - proj_hidden_dim: 4096 - proj_output_dim: 256 - pred_hidden_dim: 4096 - temperature: 0.2 -momentum: - base_tau: 0.99 - final_tau: 1.0 -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "adamw" - batch_size: 128 - lr: 3.0e-4 - classifier_lr: 3.0e-4 - weight_decay: 0.1 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1, 2, 3, 4, 5, 6, 7] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/nnclr.yaml b/solo-learn/scripts/pretrain/imagenet-100/nnclr.yaml deleted file mode 100644 index 422b7be..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/nnclr.yaml +++ /dev/null @@ -1,55 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "nnclr-imagenet100" -method: "nnclr" -backbone: - name: "resnet18" -method_kwargs: - temperature: 0.2 - proj_hidden_dim: 2048 - pred_hidden_dim: 4096 - proj_output_dim: 256 - queue_size: 65536 -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 128 - lr: 0.4 - classifier_lr: 0.1 - weight_decay: 1e-5 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/ressl.yaml b/solo-learn/scripts/pretrain/imagenet-100/ressl.yaml deleted file mode 100644 index 70416d6..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/ressl.yaml +++ /dev/null @@ -1,56 +0,0 @@ -defaults: - - _self_ - - augmentations: ressl.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "ressl-imagenet100" -method: "ressl" -backbone: - name: "resnet18" -method_kwargs: - proj_output_dim: 256 - proj_hidden_dim: 4096 - base_tau_momentum: 0.99 - final_tau_momentum: 1.0 - momentum_classifier: - temperature_q: 0.1 - temperature_k: 0.04 -momentum: - base_tau: 0.99 - final_tau: 1.0 -data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 128 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-4 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/simsiam.yaml b/solo-learn/scripts/pretrain/imagenet-100/simsiam.yaml deleted file mode 100644 index dab8055..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/simsiam.yaml +++ /dev/null @@ -1,51 +0,0 @@ -defaults: - - _self_ - - augmentations: asymmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "simsiam-imagenet100" -method: "simsiam" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 2048 - pred_hidden_dim: 512 - temperature: 0.2 -data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 128 - lr: 0.5 - classifier_lr: 0.1 - weight_decay: 1e-5 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True -augmentations_cfg: "scripts/configs/defaults/augmentations/symmetric/weak.yaml" -wandb_cfg: "scripts/configs/defaults/wandb/private.yaml" -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/supcon.yaml b/solo-learn/scripts/pretrain/imagenet-100/supcon.yaml deleted file mode 100644 index 0b91b88..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/supcon.yaml +++ /dev/null @@ -1,49 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "supcon-imagenet100" -method: "supcon" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 512 - temperature: 0.2 -data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "sgd" - batch_size: 128 - lr: 0.5 - classifier_lr: 0.1 - weight_decay: 1e-5 -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/swav.yaml b/solo-learn/scripts/pretrain/imagenet-100/swav.yaml deleted file mode 100644 index 1833f54..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/swav.yaml +++ /dev/null @@ -1,57 +0,0 @@ -defaults: - - _self_ - - augmentations: symmetric.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "swav-imagenet100" -method: "swav" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - queue_size: 3840 - proj_output_dim: 128 - num_prototypes: 3000 - epoch_queue_starts: 50 - freeze_prototypes_epochs: 2 - temperature: 0.1 -data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 128 - lr: 0.6 - classifier_lr: 0.1 - weight_decay: 1e-6 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/vibcreg.yaml b/solo-learn/scripts/pretrain/imagenet-100/vibcreg.yaml deleted file mode 100644 index ba9c891..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/vibcreg.yaml +++ /dev/null @@ -1,56 +0,0 @@ -defaults: - - _self_ - - augmentations: vicreg.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "vibcreg-imagenet100" -method: "vibcreg" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 2048 - sim_loss_weight: 25.0 - var_loss_weight: 25.0 - cov_loss_weight: 200.0 - iternorm: True -data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "lars" - batch_size: 128 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-4 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/vicreg.yaml b/solo-learn/scripts/pretrain/imagenet-100/vicreg.yaml deleted file mode 100644 index 68e817f..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/vicreg.yaml +++ /dev/null @@ -1,84 +0,0 @@ -defaults: - - _self_ - - augmentations: vicreg.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "vicreg-imagenet100" -method: "vicreg" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 2048 - proj_output_dim: 2048 - sim_loss_weight: 25.0 - var_loss_weight: 25.0 - cov_loss_weight: 1.0 -data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" - format: "dali" - num_workers: 4 -augmentations: - - rrc: - enabled: True - crop_min_scale: 0.2 - crop_max_scale: 1.0 - color_jitter: - enabled: True - brightness: 0.4 - contrast: 0.4 - saturation: 0.2 - hue: 0.1 - prob: 0.8 - grayscale: - enabled: True - prob: 0.2 - gaussian_blur: - enabled: True - prob: 0.5 - solarization: - enabled: True - prob: 0.1 - equalization: - enabled: False - prob: 0.0 - horizontal_flip: - enabled: True - prob: 0.5 - crop_size: 224 - num_crops: 2 -optimizer: - name: "lars" - batch_size: 128 - lr: 0.3 - classifier_lr: 0.1 - weight_decay: 1e-4 - kwargs: - clip_lr: True - eta: 0.02 - exclude_bias_n_norm: True -scheduler: - name: "warmup_cosine" -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet-100/wandb/mhug.yaml b/solo-learn/scripts/pretrain/imagenet-100/wandb/mhug.yaml deleted file mode 100644 index c842e44..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/wandb/mhug.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: unitn-mhug -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/imagenet-100/wandb/private.yaml b/solo-learn/scripts/pretrain/imagenet-100/wandb/private.yaml deleted file mode 100644 index ad4e200..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/wandb/private.yaml +++ /dev/null @@ -1,3 +0,0 @@ -enabled: True -entity: None -project: "gen-ssl" diff --git a/solo-learn/scripts/pretrain/imagenet-100/wmse.yaml b/solo-learn/scripts/pretrain/imagenet-100/wmse.yaml deleted file mode 100644 index 3e17478..0000000 --- a/solo-learn/scripts/pretrain/imagenet-100/wmse.yaml +++ /dev/null @@ -1,50 +0,0 @@ -defaults: - - _self_ - - augmentations: wmse.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "wmse-imagenet100" -method: "wmse" -backbone: - name: "resnet18" -method_kwargs: - proj_hidden_dim: 1024 - proj_output_dim: 64 - whitening_size: 128 -data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" - format: "dali" - num_workers: 4 -optimizer: - name: "adam" - batch_size: 128 - lr: 2e-3 - classifier_lr: 3e-3 - weight_decay: 1e-6 -scheduler: - name: "warmup_cosine" - warmup_start_lr: 0 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: [0, 1] -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet/augmentations/synthetic_symmetric.yaml b/solo-learn/scripts/pretrain/imagenet/augmentations/synthetic_symmetric.yaml index f01fed5..bde83fb 100644 --- a/solo-learn/scripts/pretrain/imagenet/augmentations/synthetic_symmetric.yaml +++ b/solo-learn/scripts/pretrain/imagenet/augmentations/synthetic_symmetric.yaml @@ -23,7 +23,7 @@ prob: 0.5 crop_size: 224 num_crops: 1 - + - rrc: enabled: True crop_min_scale: 0.08 diff --git a/solo-learn/scripts/pretrain/imagenet-100/augmentations/asymmetric.yaml b/solo-learn/scripts/pretrain/imagenet/augmentations/synthetic_symmetric_weak.yaml similarity index 66% rename from solo-learn/scripts/pretrain/imagenet-100/augmentations/asymmetric.yaml rename to solo-learn/scripts/pretrain/imagenet/augmentations/synthetic_symmetric_weak.yaml index 30d8d26..186f539 100644 --- a/solo-learn/scripts/pretrain/imagenet-100/augmentations/asymmetric.yaml +++ b/solo-learn/scripts/pretrain/imagenet/augmentations/synthetic_symmetric_weak.yaml @@ -1,3 +1,6 @@ +# Augmentations should be defined separately for synthetic and real data in the config files. That's why +# we have two lists of augmentations in defined here. The first list is for real data and the second list +# is for synthetic data. - rrc: enabled: True crop_min_scale: 0.08 @@ -6,12 +9,12 @@ prob: 0.8 brightness: 0.4 contrast: 0.4 - saturation: 0.2 + saturation: 0.4 hue: 0.1 grayscale: prob: 0.2 gaussian_blur: - prob: 1.0 + prob: 0.5 solarization: prob: 0.0 equalization: @@ -29,14 +32,14 @@ prob: 0.8 brightness: 0.4 contrast: 0.4 - saturation: 0.2 + saturation: 0.4 hue: 0.1 grayscale: prob: 0.2 gaussian_blur: - prob: 0.1 + prob: 0.5 solarization: - prob: 0.2 + prob: 0.0 equalization: prob: 0.0 horizontal_flip: diff --git a/solo-learn/scripts/pretrain/imagenet/barlow.yaml b/solo-learn/scripts/pretrain/imagenet/barlow.yaml index 1eac333..4d06ab0 100644 --- a/solo-learn/scripts/pretrain/imagenet/barlow.yaml +++ b/solo-learn/scripts/pretrain/imagenet/barlow.yaml @@ -11,7 +11,7 @@ hydra: run: dir: . -name: "barlow_twins-imagenet" +name: "barlow-imagenet" method: "barlow_twins" backbone: name: "resnet50" @@ -22,8 +22,8 @@ method_kwargs: scale_loss: 0.048 data: dataset: imagenet - train_path: "/datasets/imagenet/train" - val_path: "/datasets/imagenet/val" + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: @@ -40,10 +40,10 @@ scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: - enabled: True + enabled: False # overwrite PL stuff max_epochs: 100 diff --git a/solo-learn/scripts/pretrain/imagenet-100/byol.yaml b/solo-learn/scripts/pretrain/imagenet/barlow_diff.yaml similarity index 56% rename from solo-learn/scripts/pretrain/imagenet-100/byol.yaml rename to solo-learn/scripts/pretrain/imagenet/barlow_diff.yaml index 35cd7d5..9816a1f 100644 --- a/solo-learn/scripts/pretrain/imagenet-100/byol.yaml +++ b/solo-learn/scripts/pretrain/imagenet/barlow_diff.yaml @@ -11,45 +11,47 @@ hydra: run: dir: . -name: "byol-imagenet100" -method: "byol" +name: "barlow-imagenet-diffusion" +method: "barlow_twins" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 4096 - proj_output_dim: 256 - pred_hidden_dim: 8192 -momentum: - base_tau: 0.99 - final_tau: 1.0 + proj_output_dim: 4096 + lamb: 0.0051 + scale_loss: 0.048 data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: DIFFUSION_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 format: "dali" num_workers: 4 optimizer: name: "lars" - batch_size: 128 - lr: 0.5 + batch_size: 64 + lr: 0.8 classifier_lr: 0.1 - weight_decay: 1e-6 + weight_decay: 1.5e-6 kwargs: - clip_lr: True - eta: 0.02 + clip_lr: False + eta: 0.001 exclude_bias_n_norm: True scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: - enabled: True + enabled: False # overwrite PL stuff -max_epochs: 400 -devices: [0, 1] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/pretrain/imagenet-100-multicrop/byol.yaml b/solo-learn/scripts/pretrain/imagenet/barlow_icgan.yaml similarity index 58% rename from solo-learn/scripts/pretrain/imagenet-100-multicrop/byol.yaml rename to solo-learn/scripts/pretrain/imagenet/barlow_icgan.yaml index f17ede0..5e3f533 100644 --- a/solo-learn/scripts/pretrain/imagenet-100-multicrop/byol.yaml +++ b/solo-learn/scripts/pretrain/imagenet/barlow_icgan.yaml @@ -11,45 +11,47 @@ hydra: run: dir: . -name: "byol-multicrop-imagenet100" -method: "byol" +name: "barlow-imagenet-icgan" +method: "barlow_twins" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 4096 - proj_output_dim: 256 - pred_hidden_dim: 8192 -momentum: - base_tau: 0.99 - final_tau: 1.0 + proj_output_dim: 4096 + lamb: 0.0051 + scale_loss: 0.048 data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: ICGAN_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 format: "dali" num_workers: 4 optimizer: name: "lars" batch_size: 64 - lr: 0.5 + lr: 0.8 classifier_lr: 0.1 - weight_decay: 1e-5 + weight_decay: 1.5e-6 kwargs: - clip_lr: True - eta: 0.02 + clip_lr: False + eta: 0.001 exclude_bias_n_norm: True scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: - enabled: True + enabled: False # overwrite PL stuff -max_epochs: 400 -devices: [0, 1] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/pretrain/imagenet/byol.yaml b/solo-learn/scripts/pretrain/imagenet/byol.yaml index 54044dc..292cc59 100644 --- a/solo-learn/scripts/pretrain/imagenet/byol.yaml +++ b/solo-learn/scripts/pretrain/imagenet/byol.yaml @@ -24,16 +24,16 @@ momentum: final_tau: 1.0 data: dataset: imagenet - train_path: "/datasets/imagenet/train" - val_path: "/datasets/imagenet/val" + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: name: "lars" - batch_size: 64 - lr: 0.45 + batch_size: 256 + lr: 0.2 classifier_lr: 0.2 - weight_decay: 1e-6 + weight_decay: 15e-7 kwargs: clip_lr: False eta: 0.001 @@ -42,7 +42,7 @@ scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "/projects/imagenet_synthetic/model_checkpoints/solo-learn/solo_trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True diff --git a/solo-learn/scripts/pretrain/cifar/nnbyol.yaml b/solo-learn/scripts/pretrain/imagenet/byol_diff.yaml similarity index 61% rename from solo-learn/scripts/pretrain/cifar/nnbyol.yaml rename to solo-learn/scripts/pretrain/imagenet/byol_diff.yaml index 5cec47b..e5d8fa1 100644 --- a/solo-learn/scripts/pretrain/cifar/nnbyol.yaml +++ b/solo-learn/scripts/pretrain/imagenet/byol_diff.yaml @@ -11,47 +11,51 @@ hydra: run: dir: . -name: "nnbyol-cifar10" # change here for cifar100 -method: "nnbyol" +name: "byol-imagenet-diffusion" +method: "byol" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 4096 - pred_hidden_dim: 4096 proj_output_dim: 256 - queue_size: 65536 + pred_hidden_dim: 4096 momentum: base_tau: 0.99 final_tau: 1.0 data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: DIFFUSION_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 + format: "dali" num_workers: 4 optimizer: name: "lars" batch_size: 256 - lr: 1.0 - classifier_lr: 0.1 - weight_decay: 1e-5 + lr: 0.2 + classifier_lr: 0.2 + weight_decay: 15e-7 kwargs: - clip_lr: True - eta: 0.02 + clip_lr: False + eta: 0.001 exclude_bias_n_norm: True scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff -max_epochs: 1000 -devices: [0] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" precision: 16-mixed +accumulate_grad_batches: 16 diff --git a/solo-learn/scripts/pretrain/cifar/byol.yaml b/solo-learn/scripts/pretrain/imagenet/byol_icgan.yaml similarity index 63% rename from solo-learn/scripts/pretrain/cifar/byol.yaml rename to solo-learn/scripts/pretrain/imagenet/byol_icgan.yaml index eec6949..d7aaaff 100644 --- a/solo-learn/scripts/pretrain/cifar/byol.yaml +++ b/solo-learn/scripts/pretrain/imagenet/byol_icgan.yaml @@ -11,10 +11,10 @@ hydra: run: dir: . -name: "byol-cifar10" # change here for cifar100 +name: "byol-imagenet-icgan" method: "byol" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 4096 proj_output_dim: 256 @@ -23,34 +23,39 @@ momentum: base_tau: 0.99 final_tau: 1.0 data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: ICGAN_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 + format: "dali" num_workers: 4 optimizer: name: "lars" batch_size: 256 - lr: 1.0 - classifier_lr: 0.1 - weight_decay: 1e-5 + lr: 0.2 + classifier_lr: 0.2 + weight_decay: 15e-7 kwargs: - clip_lr: True - eta: 0.02 + clip_lr: False + eta: 0.001 exclude_bias_n_norm: True scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff -max_epochs: 1000 -devices: [0] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" precision: 16-mixed +accumulate_grad_batches: 16 diff --git a/solo-learn/scripts/pretrain/imagenet/mae.yaml b/solo-learn/scripts/pretrain/imagenet/mae.yaml deleted file mode 100644 index 7709fb3..0000000 --- a/solo-learn/scripts/pretrain/imagenet/mae.yaml +++ /dev/null @@ -1,57 +0,0 @@ -defaults: - - _self_ - - augmentations: reconstruction.yaml - - wandb: private.yaml - - override hydra/hydra_logging: disabled - - override hydra/job_logging: disabled - -# disable hydra outputs -hydra: - output_subdir: null - run: - dir: . - -name: "mae-imagenet" -method: "mae" -backbone: - name: "vit_base" -method_kwargs: - decoder_embed_dim: 512 - decoder_depth: 8 - decoder_num_heads: 16 - mask_ratio: 0.75 - norm_pix_loss: True -momentum: - base_tau: 0.9995 - final_tau: 1.0 -data: - dataset: imagenet - train_path: "/datasets/imagenet/train" - val_path: "/datasets/imagenet/val" - format: "image_folder" - num_workers: 4 -optimizer: - name: "adamw" - batch_size: 64 - lr: 2.0e-4 - classifier_lr: 2.0e-4 - weight_decay: 0.05 - kwargs: - betas: [0.9, 0.95] -scheduler: - name: "warmup_cosine" - warmup_start_lr: 0.0 -checkpoint: - enabled: True - dir: "trained_models" - frequency: 1 -auto_resume: - enabled: True - -# overwrite PL stuff -max_epochs: 400 -devices: 4 -sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet/mocov2plus.yaml b/solo-learn/scripts/pretrain/imagenet/moco.yaml similarity index 84% rename from solo-learn/scripts/pretrain/imagenet/mocov2plus.yaml rename to solo-learn/scripts/pretrain/imagenet/moco.yaml index 4ed2187..c2152a4 100644 --- a/solo-learn/scripts/pretrain/imagenet/mocov2plus.yaml +++ b/solo-learn/scripts/pretrain/imagenet/moco.yaml @@ -11,7 +11,7 @@ hydra: run: dir: . -name: "mocov2plus-imagenet" +name: "moco-imagenet" method: "mocov2plus" backbone: name: "resnet50" @@ -25,8 +25,8 @@ momentum: final_tau: 0.999 data: dataset: imagenet - train_path: "/datasets/imagenet/train" - val_path: "/datasets/imagenet/val" + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: @@ -39,10 +39,10 @@ scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: - enabled: True + enabled: False # overwrite PL stuff max_epochs: 100 diff --git a/solo-learn/scripts/pretrain/cifar/mocov2plus.yaml b/solo-learn/scripts/pretrain/imagenet/moco_diff.yaml similarity index 57% rename from solo-learn/scripts/pretrain/cifar/mocov2plus.yaml rename to solo-learn/scripts/pretrain/imagenet/moco_diff.yaml index 8c990b1..732f162 100644 --- a/solo-learn/scripts/pretrain/cifar/mocov2plus.yaml +++ b/solo-learn/scripts/pretrain/imagenet/moco_diff.yaml @@ -1,6 +1,6 @@ defaults: - _self_ - - augmentations: symmetric_weak.yaml + - augmentations: synthetic_symmetric_weak.yaml - wandb: private.yaml - override hydra/hydra_logging: disabled - override hydra/job_logging: disabled @@ -11,42 +11,46 @@ hydra: run: dir: . -name: "mocov2plus-cifar10" # change here for cifar100 +name: "moco-imagenet-diff" method: "mocov2plus" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 2048 proj_output_dim: 256 - queue_size: 32768 + queue_size: 65536 temperature: 0.2 momentum: base_tau: 0.99 final_tau: 0.999 data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: DIFFUSION_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 + format: "dali" num_workers: 4 optimizer: name: "sgd" - batch_size: 256 + batch_size: 64 lr: 0.3 - classifier_lr: 0.3 - weight_decay: 1e-4 + classifier_lr: 0.4 + weight_decay: 3e-5 scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: - enabled: True + enabled: False # overwrite PL stuff -max_epochs: 1000 -devices: [0] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/pretrain/imagenet-100/mocov2plus.yaml b/solo-learn/scripts/pretrain/imagenet/moco_icgan.yaml similarity index 61% rename from solo-learn/scripts/pretrain/imagenet-100/mocov2plus.yaml rename to solo-learn/scripts/pretrain/imagenet/moco_icgan.yaml index afbe0b4..efbf1f5 100644 --- a/solo-learn/scripts/pretrain/imagenet-100/mocov2plus.yaml +++ b/solo-learn/scripts/pretrain/imagenet/moco_icgan.yaml @@ -1,6 +1,6 @@ defaults: - _self_ - - augmentations: symmetric_weak.yaml + - augmentations: synthetic_symmetric_weak.yaml - wandb: private.yaml - override hydra/hydra_logging: disabled - override hydra/job_logging: disabled @@ -11,10 +11,10 @@ hydra: run: dir: . -name: "mocov2plus-imagenet100" +name: "moco-imagenet-icgan" method: "mocov2plus" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 2048 proj_output_dim: 256 @@ -24,29 +24,33 @@ momentum: base_tau: 0.99 final_tau: 0.999 data: - dataset: imagenet100 - train_path: "./datasets/imagenet-100/train" - val_path: "./datasets/imagenet-100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: ICGAN_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 format: "dali" num_workers: 4 optimizer: name: "sgd" - batch_size: 128 + batch_size: 64 lr: 0.3 - classifier_lr: 0.3 - weight_decay: 1e-4 + classifier_lr: 0.4 + weight_decay: 3e-5 scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: - enabled: True + enabled: False # overwrite PL stuff -max_epochs: 400 -devices: [0, 1] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/pretrain/imagenet/simclr.yaml b/solo-learn/scripts/pretrain/imagenet/simclr.yaml index f201b04..5bdd082 100644 --- a/solo-learn/scripts/pretrain/imagenet/simclr.yaml +++ b/solo-learn/scripts/pretrain/imagenet/simclr.yaml @@ -21,16 +21,16 @@ method_kwargs: temperature: 0.2 data: dataset: imagenet - train_path: "/datasets/imagenet/train" - val_path: "/datasets/imagenet/val" + train_path: TRAIN_PATH + val_path: VAL_PATH format: "dali" num_workers: 4 optimizer: name: "lars" - batch_size: 64 + batch_size: 256 lr: 0.3 classifier_lr: 0.1 - weight_decay: 1e-4 + weight_decay: 1e-6 kwargs: clip_lr: True eta: 0.02 @@ -39,7 +39,7 @@ scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "/projects/imagenet_synthetic/model_checkpoints/solo-learn/solo_trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True diff --git a/solo-learn/scripts/pretrain/imagenet-100/simclr.yaml b/solo-learn/scripts/pretrain/imagenet/simclr_diff.yaml similarity index 65% rename from solo-learn/scripts/pretrain/imagenet-100/simclr.yaml rename to solo-learn/scripts/pretrain/imagenet/simclr_diff.yaml index 8a07198..0457525 100644 --- a/solo-learn/scripts/pretrain/imagenet-100/simclr.yaml +++ b/solo-learn/scripts/pretrain/imagenet/simclr_diff.yaml @@ -1,6 +1,6 @@ defaults: - _self_ - - augmentations: symmetric.yaml + - augmentations: synthetic_symmetric.yaml - wandb: private.yaml - override hydra/hydra_logging: disabled - override hydra/job_logging: disabled @@ -11,26 +11,30 @@ hydra: run: dir: . -name: "simclr-imagenet100" +name: "simclr-imagenet-diff" method: "simclr" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 4096 proj_output_dim: 512 temperature: 0.2 data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: DIFFUSION_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 format: "dali" num_workers: 4 optimizer: name: "lars" - batch_size: 128 + batch_size: 256 lr: 0.3 classifier_lr: 0.1 - weight_decay: 1e-4 + weight_decay: 1e-6 kwargs: clip_lr: True eta: 0.02 @@ -39,14 +43,14 @@ scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff -max_epochs: 400 -devices: [0, 1] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/pretrain/imagenet-100-multicrop/simclr.yaml b/solo-learn/scripts/pretrain/imagenet/simclr_icgan.yaml similarity index 65% rename from solo-learn/scripts/pretrain/imagenet-100-multicrop/simclr.yaml rename to solo-learn/scripts/pretrain/imagenet/simclr_icgan.yaml index cbd804b..fd8b662 100644 --- a/solo-learn/scripts/pretrain/imagenet-100-multicrop/simclr.yaml +++ b/solo-learn/scripts/pretrain/imagenet/simclr_icgan.yaml @@ -1,6 +1,6 @@ defaults: - _self_ - - augmentations: symmetric.yaml + - augmentations: synthetic_symmetric.yaml - wandb: private.yaml - override hydra/hydra_logging: disabled - override hydra/job_logging: disabled @@ -11,26 +11,30 @@ hydra: run: dir: . -name: "simclr-multicrop-imagenet100" +name: "simclr-imagenet-icgan" method: "simclr" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 4096 proj_output_dim: 512 temperature: 0.2 data: - dataset: imagenet100 - train_path: "datasets/imagenet100/train" - val_path: "datasets/imagenet100/val" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: ICGAN_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 format: "dali" num_workers: 4 optimizer: name: "lars" - batch_size: 64 + batch_size: 256 lr: 0.3 classifier_lr: 0.1 - weight_decay: 1e-4 + weight_decay: 1e-6 kwargs: clip_lr: True eta: 0.02 @@ -39,14 +43,14 @@ scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff -max_epochs: 400 -devices: [0, 1] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/pretrain/cifar/mocov3.yaml b/solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_25.yaml similarity index 60% rename from solo-learn/scripts/pretrain/cifar/mocov3.yaml rename to solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_25.yaml index 9eccbd2..b4752eb 100644 --- a/solo-learn/scripts/pretrain/cifar/mocov3.yaml +++ b/solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_25.yaml @@ -1,6 +1,6 @@ defaults: - _self_ - - augmentations: asymmetric.yaml + - augmentations: synthetic_symmetric.yaml - wandb: private.yaml - override hydra/hydra_logging: disabled - override hydra/job_logging: disabled @@ -11,29 +11,29 @@ hydra: run: dir: . -name: "mocov3-cifar10" # change here for cifar100 -method: "mocov3" +name: "simclr-imagenet-diff-25" +method: "simclr" backbone: - name: "resnet18" + name: "resnet50" method_kwargs: proj_hidden_dim: 4096 - proj_output_dim: 256 - pred_hidden_dim: 4096 + proj_output_dim: 512 temperature: 0.2 -momentum: - base_tau: 0.99 - final_tau: 1.0 data: - dataset: cifar10 # change here for cifar100 - train_path: "./datasets" - val_path: "./datasets" - format: "image_folder" + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: DIFFUSION_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 0.25 + format: "dali" num_workers: 4 optimizer: name: "lars" batch_size: 256 lr: 0.3 - classifier_lr: 0.3 + classifier_lr: 0.1 weight_decay: 1e-6 kwargs: clip_lr: True @@ -43,14 +43,14 @@ scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True # overwrite PL stuff -max_epochs: 1000 -devices: [0] +max_epochs: 100 +devices: 4 sync_batchnorm: True accelerator: "gpu" strategy: "ddp" diff --git a/solo-learn/scripts/pretrain/imagenet/simclr_synthetic.yaml b/solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_50.yaml similarity index 73% rename from solo-learn/scripts/pretrain/imagenet/simclr_synthetic.yaml rename to solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_50.yaml index e1fbc2d..fefa117 100644 --- a/solo-learn/scripts/pretrain/imagenet/simclr_synthetic.yaml +++ b/solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_50.yaml @@ -11,7 +11,7 @@ hydra: run: dir: . -name: "simclr-synthetic-imagenet" +name: "simclr-imagenet-diff-50" method: "simclr" backbone: name: "resnet50" @@ -21,9 +21,9 @@ method_kwargs: temperature: 0.2 data: dataset: imagenet - train_path: "/datasets/imagenet/train" - val_path: "/datasets/imagenet/val" - synthetic_path: "/projects/imagenet_synthetic/arashaf_stablediff_batched" + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: DIFFUSION_SYNTHETIC_PATH synthetic_index_min: 0 synthetic_index_max: 9 generative_augmentation_prob: 0.5 @@ -31,10 +31,10 @@ data: num_workers: 4 optimizer: name: "lars" - batch_size: 64 + batch_size: 256 lr: 0.3 classifier_lr: 0.1 - weight_decay: 1e-4 + weight_decay: 1e-6 kwargs: clip_lr: True eta: 0.02 @@ -43,7 +43,7 @@ scheduler: name: "warmup_cosine" checkpoint: enabled: True - dir: "/projects/imagenet_synthetic/model_checkpoints/solo-learn/solo_trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True diff --git a/solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_75.yaml b/solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_75.yaml new file mode 100644 index 0000000..2e7a3bc --- /dev/null +++ b/solo-learn/scripts/pretrain/imagenet/simclr_paper_synth_stable_75.yaml @@ -0,0 +1,57 @@ +defaults: + - _self_ + - augmentations: synthetic_symmetric.yaml + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled + +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . + +name: "simclr-imagenet-diff-75" +method: "simclr" +backbone: + name: "resnet50" +method_kwargs: + proj_hidden_dim: 4096 + proj_output_dim: 512 + temperature: 0.2 +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: DIFFUSION_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 0.75 + format: "dali" + num_workers: 4 +optimizer: + name: "lars" + batch_size: 256 + lr: 0.3 + classifier_lr: 0.1 + weight_decay: 1e-6 + kwargs: + clip_lr: True + eta: 0.02 + exclude_bias_n_norm: True +scheduler: + name: "warmup_cosine" +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True + +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: "gpu" +strategy: "ddp" +precision: 16-mixed diff --git a/solo-learn/scripts/pretrain/imagenet/simsiam.yaml b/solo-learn/scripts/pretrain/imagenet/simsiam.yaml index 5439d79..1544197 100644 --- a/solo-learn/scripts/pretrain/imagenet/simsiam.yaml +++ b/solo-learn/scripts/pretrain/imagenet/simsiam.yaml @@ -4,17 +4,15 @@ defaults: - wandb: private.yaml - override hydra/hydra_logging: disabled - override hydra/job_logging: disabled - # disable hydra outputs hydra: output_subdir: null run: dir: . - -name: "simsiam-imagenet" -method: "simsiam" +name: “simsiam-imagenet” +method: “simsiam” backbone: - name: "resnet50" + name: “resnet50” method_kwargs: proj_hidden_dim: 4096 proj_output_dim: 4096 @@ -22,29 +20,28 @@ method_kwargs: temperature: 0.2 data: dataset: imagenet - train_path: "/datasets/imagenet/train" - val_path: "/datasets/imagenet/val" - format: "dali" + train_path: TRAIN_PATH + val_path: VAL_PATH + format: “dali” num_workers: 4 optimizer: - name: "sgd" + name: “sgd” batch_size: 64 lr: 0.5 classifier_lr: 0.1 weight_decay: 1e-5 scheduler: - name: "warmup_cosine" + name: “warmup_cosine” checkpoint: enabled: True - dir: "/projects/imagenet_synthetic/model_checkpoints/solo-learn/solo_trained_models" + dir: SAVE_PATH frequency: 1 auto_resume: enabled: True - # overwrite PL stuff max_epochs: 100 devices: 4 sync_batchnorm: True -accelerator: "gpu" -strategy: "ddp" -precision: 16-mixed +accelerator: “gpu” +strategy: “ddp” +precision: 16-mixed \ No newline at end of file diff --git a/solo-learn/scripts/pretrain/imagenet/simsiam_diff.yaml b/solo-learn/scripts/pretrain/imagenet/simsiam_diff.yaml new file mode 100644 index 0000000..975e3c8 --- /dev/null +++ b/solo-learn/scripts/pretrain/imagenet/simsiam_diff.yaml @@ -0,0 +1,51 @@ +defaults: + - _self_ + - augmentations: asymmetric.yaml + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . +name: “simsiam-imagenet-diff” +method: “simsiam” +backbone: + name: “resnet50” +method_kwargs: + proj_hidden_dim: 4096 + proj_output_dim: 4096 + pred_hidden_dim: 512 + temperature: 0.2 +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: DIFFUSION_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 + format: “dali” + num_workers: 4 +optimizer: + name: “sgd” + batch_size: 64 + lr: 0.5 + classifier_lr: 0.1 + weight_decay: 1e-5 +scheduler: + name: “warmup_cosine” +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: “gpu” +strategy: “ddp” +precision: 16-mixed \ No newline at end of file diff --git a/solo-learn/scripts/pretrain/imagenet/simsiam_icgan.yaml b/solo-learn/scripts/pretrain/imagenet/simsiam_icgan.yaml new file mode 100644 index 0000000..f9eacfb --- /dev/null +++ b/solo-learn/scripts/pretrain/imagenet/simsiam_icgan.yaml @@ -0,0 +1,51 @@ +defaults: + - _self_ + - augmentations: asymmetric.yaml + - wandb: private.yaml + - override hydra/hydra_logging: disabled + - override hydra/job_logging: disabled +# disable hydra outputs +hydra: + output_subdir: null + run: + dir: . +name: “simsiam-imagenet-icgan” +method: “simsiam” +backbone: + name: “resnet50” +method_kwargs: + proj_hidden_dim: 4096 + proj_output_dim: 4096 + pred_hidden_dim: 512 + temperature: 0.2 +data: + dataset: imagenet + train_path: TRAIN_PATH + val_path: VAL_PATH + synthetic_path: ICGAN_SYNTHETIC_PATH + synthetic_index_min: 0 + synthetic_index_max: 9 + generative_augmentation_prob: 1 + format: “dali” + num_workers: 4 +optimizer: + name: “sgd” + batch_size: 64 + lr: 0.5 + classifier_lr: 0.1 + weight_decay: 1e-5 +scheduler: + name: “warmup_cosine” +checkpoint: + enabled: True + dir: SAVE_PATH + frequency: 1 +auto_resume: + enabled: True +# overwrite PL stuff +max_epochs: 100 +devices: 4 +sync_batchnorm: True +accelerator: “gpu” +strategy: “ddp” +precision: 16-mixed \ No newline at end of file diff --git a/solo-learn/scripts/umap/imagenet-100/umap.sh b/solo-learn/scripts/umap/imagenet-100/umap.sh deleted file mode 100644 index 04bbb4d..0000000 --- a/solo-learn/scripts/umap/imagenet-100/umap.sh +++ /dev/null @@ -1,7 +0,0 @@ -python3 main_umap.py \ - --dataset imagenet100 \ - --train_data_path ./datasets/imagenet-100/train \ - --val_data_path ./datasets/imagenet-100/val \ - --batch_size 16 \ - --num_workers 10 \ - --pretrained_checkpoint_dir $1 From 707ced548acfab0b144a25c0c5c6aeba4f34e198 Mon Sep 17 00:00:00 2001 From: sanaAyrml Date: Thu, 16 Jan 2025 07:09:04 -0500 Subject: [PATCH 5/7] Update rest of solo --- solo-learn/.codecov.yml | 14 - solo-learn/.readthedocs.yml | 18 - solo-learn/README.md | 328 ------------------ .../downstream/object_detection/README.md | 25 -- .../configs/Base-RCNN-C4-BN.yaml | 17 - .../configs/coco_R_50_C4_2x.yaml | 13 - .../configs/coco_R_50_C4_2x_moco.yaml | 9 - .../configs/pascal_voc_R_50_C4_24k.yaml | 16 - .../configs/pascal_voc_R_50_C4_24k_moco.yaml | 9 - .../convert_model_to_detectron2.py | 55 --- solo-learn/downstream/object_detection/run.sh | 5 - .../train_object_detection.py | 81 ----- solo-learn/main_knn.py | 192 ---------- solo-learn/main_linear.py | 13 +- solo-learn/main_umap.py | 75 ---- solo-learn/zoo/cifar10.sh | 109 ------ solo-learn/zoo/cifar100.sh | 109 ------ solo-learn/zoo/imagenet.sh | 43 --- solo-learn/zoo/imagenet100.sh | 123 ------- 19 files changed, 8 insertions(+), 1246 deletions(-) delete mode 100644 solo-learn/.codecov.yml delete mode 100644 solo-learn/.readthedocs.yml delete mode 100644 solo-learn/README.md delete mode 100644 solo-learn/downstream/object_detection/README.md delete mode 100644 solo-learn/downstream/object_detection/configs/Base-RCNN-C4-BN.yaml delete mode 100644 solo-learn/downstream/object_detection/configs/coco_R_50_C4_2x.yaml delete mode 100644 solo-learn/downstream/object_detection/configs/coco_R_50_C4_2x_moco.yaml delete mode 100644 solo-learn/downstream/object_detection/configs/pascal_voc_R_50_C4_24k.yaml delete mode 100644 solo-learn/downstream/object_detection/configs/pascal_voc_R_50_C4_24k_moco.yaml delete mode 100644 solo-learn/downstream/object_detection/convert_model_to_detectron2.py delete mode 100644 solo-learn/downstream/object_detection/run.sh delete mode 100644 solo-learn/downstream/object_detection/train_object_detection.py delete mode 100644 solo-learn/main_knn.py delete mode 100644 solo-learn/main_umap.py delete mode 100644 solo-learn/zoo/cifar10.sh delete mode 100644 solo-learn/zoo/cifar100.sh delete mode 100644 solo-learn/zoo/imagenet.sh delete mode 100644 solo-learn/zoo/imagenet100.sh diff --git a/solo-learn/.codecov.yml b/solo-learn/.codecov.yml deleted file mode 100644 index 571a362..0000000 --- a/solo-learn/.codecov.yml +++ /dev/null @@ -1,14 +0,0 @@ -comment: - layout: "flags, files" - behavior: default - require_changes: false - require_base: no - require_head: no - show_carryforward_flags: true - -flag_management: - default_rules: - carryforward: true - -coverage: - range: 40...100 # custom range of coverage colors from red -> yellow -> green diff --git a/solo-learn/.readthedocs.yml b/solo-learn/.readthedocs.yml deleted file mode 100644 index b9764ee..0000000 --- a/solo-learn/.readthedocs.yml +++ /dev/null @@ -1,18 +0,0 @@ -version: 2 - -build: - os: ubuntu-20.04 - tools: - python: "3.10" - -# This part is necessary otherwise the project is not built -# Optionally set the version of Python and requirements required to build your docs -python: - install: - - requirements: docs/requirements.txt - - method: setuptools - path: . - -# By default readthedocs does not checkout git submodules -submodules: - include: all diff --git a/solo-learn/README.md b/solo-learn/README.md deleted file mode 100644 index 9e62280..0000000 --- a/solo-learn/README.md +++ /dev/null @@ -1,328 +0,0 @@ -

- -

- -
- -[![tests](https://github.com/vturrisi/solo-learn/actions/workflows/tests.yml/badge.svg)](https://github.com/vturrisi/solo-learn/actions/workflows/tests.yml) -[![Documentation Status](https://readthedocs.org/projects/solo-learn/badge/?version=latest)](https://solo-learn.readthedocs.io/en/latest/?badge=latest) -[![codecov](https://codecov.io/gh/vturrisi/solo-learn/branch/main/graph/badge.svg?token=WLU9UU17XZ)](https://codecov.io/gh/vturrisi/solo-learn) - -
- -# solo-learn -A library of self-supervised methods for unsupervised visual representation learning powered by PyTorch Lightning. -We aim at providing SOTA self-supervised methods in a comparable environment while, at the same time, implementing training tricks. -The library is self-contained, but it is possible to use the models outside of solo-learn. **More details in our [paper](#citation)**. - ---- - -## News -* **[Jan 14 2024]**: :clap: Bunch of stability improvements during 2023 :) Also added [All4One](https://openaccess.thecvf.com/content/ICCV2023/html/Estepa_All4One_Symbiotic_Neighbour_Contrastive_Learning_via_Self-Attention_and_Redundancy_Reduction_ICCV_2023_paper.html). -* **[Jan 07 2023]**: :diving_mask: Added results, checkpoints and configs for MAE on ImageNet. Thanks to [HuangChiEn](https://github.com/HuangChiEn). -* **[Dec 31 2022]**: :stars: Shiny new logo! Huge thanks to [Luiz](https://www.instagram.com/linhaaspera/)! -* **[Sep 27 2022]**: :pencil: Brand new config system using OmegaConf/Hydra. Adds more clarity and flexibility. New tutorials will follow soon! -* **[Aug 04 2022]**: :paintbrush: Added [MAE](https://arxiv.org/abs/2111.06377) and supports finetuning the backbone with `main_linear.py`, mixup, cutmix and [random augment](https://arxiv.org/abs/1909.13719). -* **[Jul 13 2022]**: :sparkling_heart: Added support for [H5](https://docs.h5py.org/en/stable/index.html) data, improved scripts and data handling. -* **[Jun 26 2022]**: :fire: Added [MoCo V3](https://arxiv.org/abs/2104.02057). -* **[Jun 10 2022]**: :bomb: Improved LARS. -* **[Jun 09 2022]**: :lollipop: Added support for [WideResnet](https://arxiv.org/abs/1605.07146), multicrop for SwAV and equalization data augmentation. -* **[May 02 2022]**: :diamond_shape_with_a_dot_inside: Wrapped Dali with a DataModule, added auto resume for linear eval and Wandb run resume. -* **[Apr 12 2022]**: :rainbow: Improved design of models and added support to train with a fraction of data. -* **[Apr 01 2022]**: :mag: Added the option to use [channel last conversion](https://pytorch.org/tutorials/intermediate/memory_format_tutorial.html#converting-existing-models) which considerably decreases training times. -* **[Feb 04 2022]**: :partying_face: Paper got accepted to JMLR. -* **[Jan 31 2022]**: :eye: Added ConvNeXt support with timm. -* **[Dec 20 2021]**: :thermometer: Added ImageNet results, scripts and checkpoints for MoCo V2+. -* **[Dec 05 2021]**: :notes: Separated [SupCon](https://arxiv.org/abs/2004.11362) from SimCLR and added runs. -* **[Dec 01 2021]**: :fountain: Added [PoolFormer](https://arxiv.org/abs/2111.11418). -* **[Nov 29 2021]**: :bangbang: Breaking changes! Update your versions!!! -* **[Nov 29 2021]**: :book: New tutorials! -* **[Nov 29 2021]**: :houses: Added offline K-NN and offline UMAP. -* **[Nov 29 2021]**: :rotating_light: Updated PyTorch and PyTorch Lightning versions. 10% faster. -* **[Nov 29 2021]**: :beers: Added code of conduct, contribution instructions, issue templates and UMAP tutorial. -* **[Nov 23 2021]**: :space_invader: Added [VIbCReg](https://arxiv.org/abs/2109.00783). -* **[Oct 21 2021]**: :triumph: Added support for object recognition via Detectron v2 and auto resume functionally that automatically tries to resume an experiment that crashed/reached a timeout. -* **[Oct 10 2021]**: :japanese_ogre: Restructured augmentation pipelines to allow more flexibility and multicrop. Also added multicrop for BYOL. -* **[Sep 27 2021]**: :pizza: Added [NNSiam](https://arxiv.org/abs/2104.14548), [NNBYOL](https://arxiv.org/abs/2104.14548), new tutorials for implementing new methods [1](https://solo-learn.readthedocs.io/en/latest/tutorials/add_new_method.html) and [2](https://solo-learn.readthedocs.io/en/latest/tutorials/add_new_method_momentum.html), more testing and fixed issues with custom data and linear evaluation. -* **[Sep 19 2021]**: :kangaroo: Added online k-NN evaluation. -* **[Sep 17 2021]**: :robot: Added [ViT](https://arxiv.org/abs/2010.11929) and [Swin](https://arxiv.org/abs/2103.14030). -* **[Sep 13 2021]**: :book: Improved [Docs](https://solo-learn.readthedocs.io/en/latest/?badge=latest) and added tutorials for [pretraining](https://solo-learn.readthedocs.io/en/latest/tutorials/overview.html) and [offline linear eval](https://solo-learn.readthedocs.io/en/latest/tutorials/offline_linear_eval.html). -* **[Aug 13 2021]**: :whale: [DeepCluster V2](https://arxiv.org/abs/2006.09882) is now available. - ---- - -## Roadmap and help needed -* Redoing the documentation to improve clarity. -* Better and up-to-date tutorials. -* Add performance-related testing to ensure that methods perform the same across updates. -* Adding new methods (continuous effort). - ---- - -## Methods available -* [All4One](https://openaccess.thecvf.com/content/ICCV2023/html/Estepa_All4One_Symbiotic_Neighbour_Contrastive_Learning_via_Self-Attention_and_Redundancy_Reduction_ICCV_2023_paper.html) -* [Barlow Twins](https://arxiv.org/abs/2103.03230) -* [BYOL](https://arxiv.org/abs/2006.07733) -* [DeepCluster V2](https://arxiv.org/abs/2006.09882) -* [DINO](https://arxiv.org/abs/2104.14294) -* [MAE](https://arxiv.org/abs/2111.06377) -* [MoCo V2+](https://arxiv.org/abs/2003.04297) -* [MoCo V3](https://arxiv.org/abs/2104.02057) -* [NNBYOL](https://arxiv.org/abs/2104.14548) -* [NNCLR](https://arxiv.org/abs/2104.14548) -* [NNSiam](https://arxiv.org/abs/2104.14548) -* [ReSSL](https://arxiv.org/abs/2107.09282) -* [SimCLR](https://arxiv.org/abs/2002.05709) -* [SimSiam](https://arxiv.org/abs/2011.10566) -* [Supervised Contrastive Learning](https://arxiv.org/abs/2004.11362) -* [SwAV](https://arxiv.org/abs/2006.09882) -* [VIbCReg](https://arxiv.org/abs/2109.00783) -* [VICReg](https://arxiv.org/abs/2105.04906) -* [W-MSE](https://arxiv.org/abs/2007.06346) - ---- - -## Extra flavor - -### Backbones -* [ResNet](https://arxiv.org/abs/1512.03385) -* [WideResNet](https://arxiv.org/abs/1605.07146) -* [ViT](https://arxiv.org/abs/2010.11929) -* [Swin](https://arxiv.org/abs/2103.14030) -* [PoolFormer](https://arxiv.org/abs/2111.11418) -* [ConvNeXt](https://arxiv.org/abs/2201.03545) - -### Data -* Increased data processing speed by up to 100% using [Nvidia Dali](https://github.com/NVIDIA/DALI). -* Flexible augmentations. - -### Evaluation -* Online linear evaluation via stop-gradient for easier debugging and prototyping (optionally available for the momentum backbone as well). -* Standard offline linear evaluation. -* Online and offline K-NN evaluation. -* Automatic feature space visualization with UMAP. - -### Training tricks -* All the perks of PyTorch Lightning (mixed precision, gradient accumulation, clipping, and much more). -* Channel last conversion -* Multi-cropping dataloading following [SwAV](https://arxiv.org/abs/2006.09882): - * **Note**: currently, only SimCLR, BYOL and SwAV support this. -* Exclude batchnorm and biases from weight decay and LARS. -* No LR scheduler for the projection head (as in SimSiam). - -### Logging -* Metric logging on the cloud with [WandB](https://wandb.ai/site) -* Custom model checkpointing with a simple file organization. - ---- -## Requirements -* torch -* torchvision -* tqdm -* einops -* wandb -* pytorch-lightning -* lightning-bolts -* torchmetrics -* scipy -* timm - -**Optional**: -* nvidia-dali -* matplotlib -* seaborn -* pandas -* umap-learn - ---- - -## Installation - -First clone the repo. - -Then, to install solo-learn with [Dali](https://github.com/NVIDIA/DALI) and/or UMAP support, use: -```bash -pip3 install .[dali,umap,h5] --extra-index-url https://developer.download.nvidia.com/compute/redist -``` - -If no Dali/UMAP/H5 support is needed, the repository can be installed as: -```bash -pip3 install . -``` - -For local development: -```bash -pip3 install -e .[umap,h5] -# Make sure you have pre-commit hooks installed -pre-commit install -``` - -**NOTE:** if you are having trouble with dali, install it following their [guide](https://github.com/NVIDIA/DALI). - -**NOTE 2:** consider installing [Pillow-SIMD](https://github.com/uploadcare/pillow-simd) for better loading times when not using Dali. - -**NOTE 3:** Soon to be on pip. - ---- - -## Training - -For pretraining the backbone, follow one of the many bash files in `scripts/pretrain/`. -We are now using [Hydra](https://github.com/facebookresearch/hydra) to handle the config files, so the common syntax is something like: -```bash -python3 main_pretrain.py \ - # path to training script folder - --config-path scripts/pretrain/imagenet-100/ \ - # training config name - --config-name barlow.yaml - # add new arguments (e.g. those not defined in the yaml files) - # by doing ++new_argument=VALUE - # pytorch lightning's arguments can be added here as well. -``` - -After that, for offline linear evaluation, follow the examples in `scripts/linear` or `scripts/finetune` for finetuning the whole backbone. - -For k-NN evaluation and UMAP visualization check the scripts in `scripts/{knn,umap}`. - -**NOTE:** Files try to be up-to-date and follow as closely as possible the recommended parameters of each paper, but check them before running. - ---- - -## Tutorials - -Please, check out our [documentation](https://solo-learn.readthedocs.io/en/latest) and tutorials: -* [Overview](https://solo-learn.readthedocs.io/en/latest/tutorials/overview.html) -* [Offline linear eval](https://solo-learn.readthedocs.io/en/latest/tutorials/offline_linear_eval.html) -* [Object detection](https://github.com/vturrisi/solo-learn/blob/main/downstream/object_detection/README.md) -* [Adding a new method](https://github.com/vturrisi/solo-learn/blob/main/docs/source/tutorials/add_new_method.rst) -* [Adding a new momentum method](https://github.com/vturrisi/solo-learn/blob/main/docs/source/tutorials/add_new_method_momentum.rst) -* [Visualizing features with UMAP](https://github.com/vturrisi/solo-learn/blob/main/docs/source/tutorials/umap.rst) -* [Offline k-NN](https://github.com/vturrisi/solo-learn/blob/main/docs/source/tutorials/knn.rst) - -If you want to contribute to solo-learn, make sure you take a look at [how to contribute](https://github.com/vturrisi/solo-learn/blob/main/.github/CONTRIBUTING.md) and follow the [code of conduct](https://github.com/vturrisi/solo-learn/blob/main/.github/CODE_OF_CONDUCT.md) - ---- - -## Model Zoo - -All pretrained models avaiable can be downloaded directly via the tables below or programmatically by running one of the following scripts -`zoo/cifar10.sh`, `zoo/cifar100.sh`, `zoo/imagenet100.sh` and `zoo/imagenet.sh`. - ---- - -## Results - -**Note:** hyperparameters may not be the best, we will be re-running the methods with lower performance eventually. - -### CIFAR-10 - -| Method | Backbone | Epochs | Dali | Acc@1 | Acc@5 | Checkpoint | -|--------------|:--------:|:------:|:----:|:--------------:|:--------------:|:----------:| -| All4One | ResNet18 | 1000 | :x: | 93.24 | 99.88 | [:link:](https://drive.google.com/drive/folders/1dtYmZiftruQ7B2PQ8fo44wguCZ0eSzAd?usp=sharing) | -| Barlow Twins | ResNet18 | 1000 | :x: | 92.10 | 99.73 | [:link:](https://drive.google.com/drive/folders/1L5RAM3lCSViD2zEqLtC-GQKVw6mxtxJ_?usp=sharing) | -| BYOL | ResNet18 | 1000 | :x: | 92.58 | 99.79 | [:link:](https://drive.google.com/drive/folders/1KxeYAEE7Ev9kdFFhXWkPZhG-ya3_UwGP?usp=sharing) | -|DeepCluster V2| ResNet18 | 1000 | :x: | 88.85 | 99.58 | [:link:](https://drive.google.com/drive/folders/1tkEbiDQ38vZaQUsT6_vEpxbDxSUAGwF-?usp=sharing) | -| DINO | ResNet18 | 1000 | :x: | 89.52 | 99.71 | [:link:](https://drive.google.com/drive/folders/1vyqZKUyP8sQyEyf2cqonxlGMbQC-D1Gi?usp=sharing) | -| MoCo V2+ | ResNet18 | 1000 | :x: | 92.94 | 99.79 | [:link:](https://drive.google.com/drive/folders/1ruNFEB3F-Otxv2Y0p62wrjA4v5Fr2cKC?usp=sharing) | -| MoCo V3 | ResNet18 | 1000 | :x: | 93.10 | 99.80 | [:link:](https://drive.google.com/drive/folders/1KwZTshNEpmqnYJcmyYPvfIJ_DNwqtAVj?usp=sharing) | -| NNCLR | ResNet18 | 1000 | :x: | 91.88 | 99.78 | [:link:](https://drive.google.com/drive/folders/1xdCzhvRehPmxinphuiZqFlfBwfwWDcLh?usp=sharing) | -| ReSSL | ResNet18 | 1000 | :x: | 90.63 | 99.62 | [:link:](https://drive.google.com/drive/folders/1jrFcztY2eO_fG98xPshqOD15pDIhLXp-?usp=sharing) | -| SimCLR | ResNet18 | 1000 | :x: | 90.74 | 99.75 | [:link:](https://drive.google.com/drive/folders/1mcvWr8P2WNJZ7TVpdLHA_Q91q4VK3y8O?usp=sharing) | -| Simsiam | ResNet18 | 1000 | :x: | 90.51 | 99.72 | [:link:](https://drive.google.com/drive/folders/1OO_igM3IK5oDw7GjQTNmdfg2I1DH3xOk?usp=sharing) | -| SupCon | ResNet18 | 1000 | :x: | 93.82 | 99.65 | [:link:](https://drive.google.com/drive/folders/1VwZ9TrJXCpnxyo7P_l397yGrGH-DAUv-?usp=sharing) | -| SwAV | ResNet18 | 1000 | :x: | 89.17 | 99.68 | [:link:](https://drive.google.com/drive/folders/1nlJH4Ljm8-5fOIeAaKppQT6gtsmmW1T0?usp=sharing) | -| VIbCReg | ResNet18 | 1000 | :x: | 91.18 | 99.74 | [:link:](https://drive.google.com/drive/folders/1XvxUOnLPZlC_-OkeuO7VqXT7z9_tNVk7?usp=sharing) | -| VICReg | ResNet18 | 1000 | :x: | 92.07 | 99.74 | [:link:](https://drive.google.com/drive/folders/159ZgCxocB7aaHxwNDubnAWU71zXV9hn-?usp=sharing) | -| W-MSE | ResNet18 | 1000 | :x: | 88.67 | 99.68 | [:link:](https://drive.google.com/drive/folders/1xPCiULzQ4JCmhrTsbxBp9S2jRZ01KiVM?usp=sharing) | - - -### CIFAR-100 - -| Method | Backbone | Epochs | Dali | Acc@1 | Acc@5 | Checkpoint | -|--------------|:--------:|:------:|:----:|:--------------:|:--------------:|:----------:| -| All4One | ResNet18 | 1000 | :x: | 72.17 | 93.35 | [:link:](https://drive.google.com/drive/folders/1oQcC80XPr-Wxhjs-PEqD_8VhUa_izqeZ?usp=sharing) | -| Barlow Twins | ResNet18 | 1000 | :x: | 70.90 | 91.91 | [:link:](https://drive.google.com/drive/folders/1hDLSApF3zSMAKco1Ck4DMjyNxhsIR2yq?usp=sharing) | -| BYOL | ResNet18 | 1000 | :x: | 70.46 | 91.96 | [:link:](https://drive.google.com/drive/folders/1hwsEdsfsUulD2tAwa4epKK9pkSuvFv6m?usp=sharing) | -|DeepCluster V2| ResNet18 | 1000 | :x: | 63.61 | 88.09 | [:link:](https://drive.google.com/drive/folders/1gAKyMz41mvGh1BBOYdc_xu6JPSkKlWqK?usp=sharing) | -| DINO | ResNet18 | 1000 | :x: | 66.76 | 90.34 | [:link:](https://drive.google.com/drive/folders/1TxeZi2YLprDDtbt_y5m29t4euroWr1Fy?usp=sharing) | -| MoCo V2+ | ResNet18 | 1000 | :x: | 69.89 | 91.65 | [:link:](https://drive.google.com/drive/folders/15oWNM16vO6YVYmk_yOmw2XUrFivRXam4?usp=sharing) | -| MoCo V3 | ResNet18 | 1000 | :x: | 68.83 | 90.57 | [:link:](https://drive.google.com/drive/folders/1Hcf9kMIADKydfxvXLquY9nv7sfNaJ3v6?usp=sharing) | -| NNCLR | ResNet18 | 1000 | :x: | 69.62 | 91.52 | [:link:](https://drive.google.com/drive/folders/1Dz72o0-5hugYPW1kCCQDBb0Xi3kzMLzu?usp=sharing) | -| ReSSL | ResNet18 | 1000 | :x: | 65.92 | 89.73 | [:link:](https://drive.google.com/drive/folders/1aVZs9cHAu6Ccz8ILyWkp6NhTsJGBGfjr?usp=sharing) | -| SimCLR | ResNet18 | 1000 | :x: | 65.78 | 89.04 | [:link:](https://drive.google.com/drive/folders/13pGPcOO9Y3rBoeRVWARgbMFEp8OXxZa0?usp=sharing) | -| Simsiam | ResNet18 | 1000 | :x: | 66.04 | 89.62 | [:link:](https://drive.google.com/drive/folders/1AJUPmsIHh_nqEcFe-Vcz2o4ruEibFHWO?usp=sharing) | -| SupCon | ResNet18 | 1000 | :x: | 70.38 | 89.57 | [:link:](https://drive.google.com/drive/folders/15C68oHPDMAOPtmBAm_Xw6YI6GgOW00gM?usp=sharing) | -| SwAV | ResNet18 | 1000 | :x: | 64.88 | 88.78 | [:link:](https://drive.google.com/drive/folders/1U_bmyhlPEN941hbx0SdRGOT4ivCarQB9?usp=sharing) | -| VIbCReg | ResNet18 | 1000 | :x: | 67.37 | 90.07 | [:link:](https://drive.google.com/drive/folders/19u3p1maX3xqwoCHNrqSDb98J5fRvd_6v?usp=sharing) | -| VICReg | ResNet18 | 1000 | :x: | 68.54 | 90.83 | [:link:](https://drive.google.com/drive/folders/1AHmVf_Zl5fikkmR4X3NWlmMOnRzfv0aT?usp=sharing) | -| W-MSE | ResNet18 | 1000 | :x: | 61.33 | 87.26 | [:link:](https://drive.google.com/drive/folders/1vc9j3RLpVCbECh6o-44oMiE5snNyKPlF?usp=sharing) | - -### ImageNet-100 - -| Method | Backbone | Epochs | Dali | Acc@1 (online) | Acc@1 (offline) | Acc@5 (online) | Acc@5 (offline) | Checkpoint | -|-------------------------|:--------:|:------:|:------------------:|:--------------:|:---------------:|:--------------:|:---------------:|:----------:| -| All4One | ResNet18 | 400 | :heavy_check_mark: | 81.93 | - | 96.23 | - | [:link:](https://drive.google.com/drive/folders/1bJCRLP5Rz_JEylNq9C4sY3ccYZSchUGR?usp=sharing) | -| Barlow Twins :rocket: | ResNet18 | 400 | :heavy_check_mark: | 80.38 | 80.16 | 95.28 | 95.14 | [:link:](https://drive.google.com/drive/folders/1rj8RbER9E71mBlCHIZEIhKPUFn437D5O?usp=sharing) | -| BYOL :rocket: | ResNet18 | 400 | :heavy_check_mark: | 80.16 | 80.32 | 95.02 | 94.94 | [:link:](https://drive.google.com/drive/folders/1riOLjMawD_znO4HYj8LBN2e1X4jXpDE1?usp=sharing) | -| DeepCluster V2 | ResNet18 | 400 | :x: | 75.36 | 75.4 | 93.22 | 93.10 | [:link:](https://drive.google.com/drive/folders/1d5jPuavrQ7lMlQZn5m2KnN5sPMGhHFo8?usp=sharing) | -| DINO | ResNet18 | 400 | :heavy_check_mark: | 74.84 | 74.92 | 92.92 | 92.78 | [:link:](https://drive.google.com/drive/folders/1NtVvRj-tQJvrMxRlMtCJSAecQnYZYkqs?usp=sharing) | -| DINO :sleepy: | ViT Tiny | 400 | :x: | 63.04 | TODO | 87.72 | TODO | [:link:](https://drive.google.com/drive/folders/16AfsM-UpKky43kdSMlqj4XRe69pRdJLc?usp=sharing) | -| MoCo V2+ :rocket: | ResNet18 | 400 | :heavy_check_mark: | 78.20 | 79.28 | 95.50 | 95.18 | [:link:](https://drive.google.com/drive/folders/1ItYBtMJ23Yh-Rhrvwjm4w1waFfUGSoKX?usp=sharing) | -| MoCo V3 :rocket: | ResNet18 | 400 | :heavy_check_mark: | 80.36 | 80.36 | 95.18 | 94.96 | [:link:](https://drive.google.com/drive/folders/15J0JiZsQAsrQler8mbbio-desb_nVoD1?usp=sharing) | -| MoCo V3 :rocket: | ResNet50 | 400 | :heavy_check_mark: | 85.48 | 84.58 | 96.82 | 96.70 | [:link:](https://drive.google.com/drive/folders/1a1VRXGlP50COZ57DPUA_doBmpaxGKpQE?usp=sharing) | -| NNCLR :rocket: | ResNet18 | 400 | :heavy_check_mark: | 79.80 | 80.16 | 95.28 | 95.30 | [:link:](https://drive.google.com/drive/folders/1QMkq8w3UsdcZmoNUIUPgfSCAZl_LSNjZ?usp=sharing) | -| ReSSL | ResNet18 | 400 | :heavy_check_mark: | 76.92 | 78.48 | 94.20 | 94.24 | [:link:](https://drive.google.com/drive/folders/1urWIFACLont4GAduis6l0jcEbl080c9U?usp=sharing) | -| SimCLR :rocket: | ResNet18 | 400 | :heavy_check_mark: | 77.64 | TODO | 94.06 | TODO | [:link:](https://drive.google.com/drive/folders/1yxAVKnc8Vf0tDfkixSB5mXe7dsA8Ll37?usp=sharing) | -| Simsiam | ResNet18 | 400 | :heavy_check_mark: | 74.54 | 78.72 | 93.16 | 94.78 | [:link:](https://drive.google.com/drive/folders/1Bc8Xj-Z7ILmspsiEQHyQsTOn4M99F_f5?usp=sharing) | -| SupCon | ResNet18 | 400 | :heavy_check_mark: | 84.40 | TODO | 95.72 | TODO | [:link:](https://drive.google.com/drive/folders/1BzR0nehkCKpnLhi-oeDynzzUcCYOCUJi?usp=sharing) | -| SwAV | ResNet18 | 400 | :heavy_check_mark: | 74.04 | 74.28 | 92.70 | 92.84 | [:link:](https://drive.google.com/drive/folders/1VWCMM69sokzjVoPzPSLIsUy5S2Rrm1xJ?usp=sharing) | -| VIbCReg | ResNet18 | 400 | :heavy_check_mark: | 79.86 | 79.38 | 94.98 | 94.60 | [:link:](https://drive.google.com/drive/folders/1Q06hH18usvRwj2P0bsmoCkjNUX_0syCK?usp=sharing) | -| VICReg :rocket: | ResNet18 | 400 | :heavy_check_mark: | 79.22 | 79.40 | 95.06 | 95.02 | [:link:](https://drive.google.com/drive/folders/1uWWR5VBUru8vaHaGeLicS6X3R4CfZsr2?usp=sharing) | -| W-MSE | ResNet18 | 400 | :heavy_check_mark: | 67.60 | 69.06 | 90.94 | 91.22 | [:link:](https://drive.google.com/drive/folders/1TxubagNV4z5Qs7SqbBcyRHWGKevtFO5l?usp=sharing) | - -:rocket: methods where hyperparameters were heavily tuned. - -:sleepy: ViT is very compute intensive and unstable, so we are slowly running larger architectures and with a larger batch size. Atm, total batch size is 128 and we needed to use float32 precision. If you want to contribute by running it, let us know! - -### ImageNet - -| Method | Backbone | Epochs | Dali | Acc@1 (online) | Acc@1 (offline) | Acc@5 (online) | Acc@5 (offline) | Checkpoint | Finetuned Checkpoint -|--------------|:--------:|:------:|:------------------:|:--------------:|:---------------:|:--------------:|:---------------:|:----------:|:----------:| -| Barlow Twins | ResNet50 | 100 | :heavy_check_mark: | 67.18 | 67.23 | 87.69 | 87.98 | [:link:](https://drive.google.com/drive/folders/1IQUIrCOSduAjUJ31WJ1G5tHDZzWUIEft?usp=sharing) | | -| BYOL | ResNet50 | 100 | :heavy_check_mark: | 68.63 | 68.37 | 88.80 | 88.66 | [:link:](https://drive.google.com/drive/folders/1-UXo-MttdrqiEQXfV4Duc93fA3mIdsha?usp=sharing) | | -| MoCo V2+ | ResNet50 | 100 | :heavy_check_mark: | 62.61 | 66.84 | 85.40 | 87.60 | [:link:](https://drive.google.com/drive/folders/1NiBDmieEpNqkwrgn_H7bMnEDVAYc8Sk7?usp=sharing) | | -| MAE | ViT-B/16 | 100 | :x: | ~ | 81.60 (finetuned) | ~ | 95.50 (finetuned) | [:link:](https://drive.google.com/drive/folders/1OuaXCnQ7WeqyKPxfJibAkXoVTx7S8Hbb) | [:link:](https://drive.google.com/drive/folders/1c9DGhmLsTTtOu2vc9rodqm89wKtp40C5) | - - - -## Training efficiency for DALI - -We report the training efficiency of some methods using a ResNet18 with and without DALI (4 workers per GPU) in a server with an Intel i9-9820X and two RTX2080ti. - -| Method | Dali | Total time for 20 epochs | Time for 1 epoch | GPU memory (per GPU) | -|--------------|:----------------:|:--------------------------:|:--------------------:|:---------------------:| -| Barlow Twins | :x: | 1h 38m 27s | 4m 55s | 5097 MB | -| |:heavy_check_mark:| 43m 2s | 2m 10s (56% faster) | 9292 MB | -| BYOL | :x: | 1h 38m 46s | 4m 56s | 5409 MB | -| |:heavy_check_mark:| 50m 33s | 2m 31s (49% faster) | 9521 MB | -| NNCLR | :x: | 1h 38m 30s | 4m 55s | 5060 MB | -| |:heavy_check_mark:| 42m 3s | 2m 6s (64% faster) | 9244 MB | - -**Note**: GPU memory increase doesn't scale with the model, rather it scales with the number of workers. - ---- - -## Citation -If you use solo-learn, please cite our [paper](https://jmlr.org/papers/v23/21-1155.html): -```bibtex -@article{JMLR:v23:21-1155, - author = {Victor Guilherme Turrisi da Costa and Enrico Fini and Moin Nabi and Nicu Sebe and Elisa Ricci}, - title = {solo-learn: A Library of Self-supervised Methods for Visual Representation Learning}, - journal = {Journal of Machine Learning Research}, - year = {2022}, - volume = {23}, - number = {56}, - pages = {1-6}, - url = {http://jmlr.org/papers/v23/21-1155.html} -} -``` diff --git a/solo-learn/downstream/object_detection/README.md b/solo-learn/downstream/object_detection/README.md deleted file mode 100644 index 0081a19..0000000 --- a/solo-learn/downstream/object_detection/README.md +++ /dev/null @@ -1,25 +0,0 @@ - - - -## Transferring to Detection - -The `train_object_detection.py` script reproduces the object detection experiments on Pascal VOC and COCO. - -### Instruction - -1. Install [detectron2](https://github.com/facebookresearch/detectron2/blob/master/INSTALL.md). - -1. Convert a pre-trained model to detectron2's format: - ``` - python3 convert_model_to_detectron2.py --pretrained_feature_extractor PATH_TO_CKPT --output_detectron_model ./detectron_model.pkl - ``` - -1. Put dataset under "./datasets" directory, - following the [directory structure](https://github.com/facebookresearch/detectron2/tree/master/datasets) - requried by detectron2. - -1. Run training: - ``` - python train_net.py --config-file configs/pascal_voc_R_50_C4_24k_moco.yaml \ - --num-gpus 8 MODEL.WEIGHTS ./detectron_model.pkl - ``` diff --git a/solo-learn/downstream/object_detection/configs/Base-RCNN-C4-BN.yaml b/solo-learn/downstream/object_detection/configs/Base-RCNN-C4-BN.yaml deleted file mode 100644 index 5104c6a..0000000 --- a/solo-learn/downstream/object_detection/configs/Base-RCNN-C4-BN.yaml +++ /dev/null @@ -1,17 +0,0 @@ -MODEL: - META_ARCHITECTURE: "GeneralizedRCNN" - RPN: - PRE_NMS_TOPK_TEST: 6000 - POST_NMS_TOPK_TEST: 1000 - ROI_HEADS: - NAME: "Res5ROIHeadsExtraNorm" - BACKBONE: - FREEZE_AT: 0 - RESNETS: - NORM: "SyncBN" -TEST: - PRECISE_BN: - ENABLED: True -SOLVER: - IMS_PER_BATCH: 16 - BASE_LR: 0.02 diff --git a/solo-learn/downstream/object_detection/configs/coco_R_50_C4_2x.yaml b/solo-learn/downstream/object_detection/configs/coco_R_50_C4_2x.yaml deleted file mode 100644 index 5b7e424..0000000 --- a/solo-learn/downstream/object_detection/configs/coco_R_50_C4_2x.yaml +++ /dev/null @@ -1,13 +0,0 @@ -_BASE_: "Base-RCNN-C4-BN.yaml" -MODEL: - MASK_ON: True - WEIGHTS: "detectron2://ImageNetPretrained/MSRA/R-50.pkl" -INPUT: - MIN_SIZE_TRAIN: (640, 672, 704, 736, 768, 800) - MIN_SIZE_TEST: 800 -DATASETS: - TRAIN: ("coco_2017_train",) - TEST: ("coco_2017_val",) -SOLVER: - STEPS: (120000, 160000) - MAX_ITER: 180000 diff --git a/solo-learn/downstream/object_detection/configs/coco_R_50_C4_2x_moco.yaml b/solo-learn/downstream/object_detection/configs/coco_R_50_C4_2x_moco.yaml deleted file mode 100644 index 73ef270..0000000 --- a/solo-learn/downstream/object_detection/configs/coco_R_50_C4_2x_moco.yaml +++ /dev/null @@ -1,9 +0,0 @@ -_BASE_: "coco_R_50_C4_2x.yaml" -MODEL: - PIXEL_MEAN: [123.675, 116.280, 103.530] - PIXEL_STD: [58.395, 57.120, 57.375] - WEIGHTS: "See Instructions" - RESNETS: - STRIDE_IN_1X1: False -INPUT: - FORMAT: "RGB" diff --git a/solo-learn/downstream/object_detection/configs/pascal_voc_R_50_C4_24k.yaml b/solo-learn/downstream/object_detection/configs/pascal_voc_R_50_C4_24k.yaml deleted file mode 100644 index a05eb5e..0000000 --- a/solo-learn/downstream/object_detection/configs/pascal_voc_R_50_C4_24k.yaml +++ /dev/null @@ -1,16 +0,0 @@ -_BASE_: "Base-RCNN-C4-BN.yaml" -MODEL: - MASK_ON: False - WEIGHTS: "detectron2://ImageNetPretrained/MSRA/R-50.pkl" - ROI_HEADS: - NUM_CLASSES: 20 -INPUT: - MIN_SIZE_TRAIN: (480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800) - MIN_SIZE_TEST: 800 -DATASETS: - TRAIN: ('voc_2007_trainval', 'voc_2012_trainval') - TEST: ('voc_2007_test',) -SOLVER: - STEPS: (18000, 22000) - MAX_ITER: 24000 - WARMUP_ITERS: 100 diff --git a/solo-learn/downstream/object_detection/configs/pascal_voc_R_50_C4_24k_moco.yaml b/solo-learn/downstream/object_detection/configs/pascal_voc_R_50_C4_24k_moco.yaml deleted file mode 100644 index eebe690..0000000 --- a/solo-learn/downstream/object_detection/configs/pascal_voc_R_50_C4_24k_moco.yaml +++ /dev/null @@ -1,9 +0,0 @@ -_BASE_: "pascal_voc_R_50_C4_24k.yaml" -MODEL: - PIXEL_MEAN: [123.675, 116.280, 103.530] - PIXEL_STD: [58.395, 57.120, 57.375] - WEIGHTS: "See Instructions" - RESNETS: - STRIDE_IN_1X1: False -INPUT: - FORMAT: "RGB" diff --git a/solo-learn/downstream/object_detection/convert_model_to_detectron2.py b/solo-learn/downstream/object_detection/convert_model_to_detectron2.py deleted file mode 100644 index aa977ad..0000000 --- a/solo-learn/downstream/object_detection/convert_model_to_detectron2.py +++ /dev/null @@ -1,55 +0,0 @@ -# Copyright 2021 solo-learn development team. - -# Permission is hereby granted, free of charge, to any person obtaining a copy of -# this software and associated documentation files (the "Software"), to deal in -# the Software without restriction, including without limitation the rights to use, -# copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the -# Software, and to permit persons to whom the Software is furnished to do so, -# subject to the following conditions: - -# The above copyright notice and this permission notice shall be included in all copies -# or substantial portions of the Software. - -# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, -# INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR -# PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE -# FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR -# OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER -# DEALINGS IN THE SOFTWARE. - -from argparse import ArgumentParser -import pickle as pkl -import torch - -if __name__ == "__main__": - parser = ArgumentParser() - parser.add_argument("--pretrained_feature_extractor", type=str, required=True) - parser.add_argument("--output_detectron_model", type=str, required=True) - - args = parser.parse_args() - - checkpoint = torch.load(args.pretrained_feature_extractor, map_location="cpu") - checkpoint = checkpoint["state_dict"] - - newmodel = {} - for k, v in checkpoint.items(): - if not k.startswith("backbone"): - continue - - old_k = k - k = k.replace("backbone.", "") - if "layer" not in k: - k = "stem." + k - for t in [1, 2, 3, 4]: - k = k.replace(f"layer{t}", f"res{t + 1}") - for t in [1, 2, 3]: - k = k.replace(f"bn{t}", f"conv{t}.norm") - k = k.replace("downsample.0", "shortcut") - k = k.replace("downsample.1", "shortcut.norm") - print(old_k, "->", k) - newmodel[k] = v.numpy() - - res = {"model": newmodel, "__author__": "solo-learn", "matching_heuristics": True} - - with open(args.output_detectron_model, "wb") as f: - pkl.dump(res, f) diff --git a/solo-learn/downstream/object_detection/run.sh b/solo-learn/downstream/object_detection/run.sh deleted file mode 100644 index a3bc616..0000000 --- a/solo-learn/downstream/object_detection/run.sh +++ /dev/null @@ -1,5 +0,0 @@ -export DETECTRON2_DATASETS=/data/datasets - -# good results for BYOL -python3 train_object_detection.py --config-file configs/pascal_voc_R_50_C4_24k_moco.yaml \ - --num-gpus 2 MODEL.WEIGHTS ./detectron_model.pkl SOLVER.IMS_PER_BATCH 16 SOLVER.BASE_LR 0.1 diff --git a/solo-learn/downstream/object_detection/train_object_detection.py b/solo-learn/downstream/object_detection/train_object_detection.py deleted file mode 100644 index 856955f..0000000 --- a/solo-learn/downstream/object_detection/train_object_detection.py +++ /dev/null @@ -1,81 +0,0 @@ -# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved -# Copied from https://github.com/facebookresearch/moco/blob/main/detection/train_net.py - -import os - -from detectron2.checkpoint import DetectionCheckpointer -from detectron2.config import get_cfg -from detectron2.engine import ( - DefaultTrainer, - default_argument_parser, - default_setup, - launch, -) -from detectron2.evaluation import COCOEvaluator, PascalVOCDetectionEvaluator -from detectron2.layers import get_norm -from detectron2.modeling.roi_heads import ROI_HEADS_REGISTRY, Res5ROIHeads - - -@ROI_HEADS_REGISTRY.register() -class Res5ROIHeadsExtraNorm(Res5ROIHeads): - """ - As described in the MOCO paper, there is an extra BN layer - following the res5 stage. - """ - - def _build_res5_block(self, cfg): - seq, out_channels = super()._build_res5_block(cfg) - norm = cfg.MODEL.RESNETS.NORM - norm = get_norm(norm, out_channels) - seq.add_module("norm", norm) - return seq, out_channels - - -class Trainer(DefaultTrainer): - @classmethod - def build_evaluator(cls, cfg, dataset_name, output_folder=None): - if output_folder is None: - output_folder = os.path.join(cfg.OUTPUT_DIR, "inference") - if "coco" in dataset_name: - return COCOEvaluator(dataset_name, cfg, True, output_folder) - else: - assert "voc" in dataset_name - return PascalVOCDetectionEvaluator(dataset_name) - - -def setup(args): - cfg = get_cfg() - cfg.merge_from_file(args.config_file) - cfg.merge_from_list(args.opts) - cfg.freeze() - default_setup(cfg, args) - return cfg - - -def main(args): - cfg = setup(args) - - if args.eval_only: - model = Trainer.build_model(cfg) - DetectionCheckpointer(model, save_dir=cfg.OUTPUT_DIR).resume_or_load( - cfg.MODEL.WEIGHTS, resume=args.resume - ) - res = Trainer.test(cfg, model) - return res - - trainer = Trainer(cfg) - trainer.resume_or_load(resume=args.resume) - return trainer.train() - - -if __name__ == "__main__": - args = default_argument_parser().parse_args() - print("Command Line Args:", args) - launch( - main, - args.num_gpus, - num_machines=args.num_machines, - machine_rank=args.machine_rank, - dist_url=args.dist_url, - args=(args,), - ) diff --git a/solo-learn/main_knn.py b/solo-learn/main_knn.py deleted file mode 100644 index a8a39a0..0000000 --- a/solo-learn/main_knn.py +++ /dev/null @@ -1,192 +0,0 @@ -# Copyright 2023 solo-learn development team. - -# Permission is hereby granted, free of charge, to any person obtaining a copy of -# this software and associated documentation files (the "Software"), to deal in -# the Software without restriction, including without limitation the rights to use, -# copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the -# Software, and to permit persons to whom the Software is furnished to do so, -# subject to the following conditions: - -# The above copyright notice and this permission notice shall be included in all copies -# or substantial portions of the Software. - -# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, -# INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR -# PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE -# FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR -# OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER -# DEALINGS IN THE SOFTWARE. - -import json -import os -from pathlib import Path -from typing import Tuple - -import torch -import torch.nn as nn -from omegaconf import OmegaConf -from torch.utils.data import DataLoader -from tqdm import tqdm - -from solo.args.knn import parse_args_knn -from solo.data.classification_dataloader import ( - prepare_dataloaders, - prepare_datasets, - prepare_transforms, -) -from solo.methods import METHODS -from solo.utils.knn import WeightedKNNClassifier - - -@torch.no_grad() -def extract_features(loader: DataLoader, model: nn.Module) -> Tuple[torch.Tensor]: - """Extract features from a data loader using a model. - - Args: - loader (DataLoader): dataloader for a dataset. - model (nn.Module): torch module used to extract features. - - Returns: - Tuple(torch.Tensor): tuple containing the backbone features, projector features and labels. - """ - - model.eval() - backbone_features, proj_features, labels = [], [], [] - for im, lab in tqdm(loader): - im = im.cuda(non_blocking=True) - lab = lab.cuda(non_blocking=True) - outs = model(im) - backbone_features.append(outs["feats"].detach()) - proj_features.append(outs["z"]) - labels.append(lab) - model.train() - backbone_features = torch.cat(backbone_features) - proj_features = torch.cat(proj_features) - labels = torch.cat(labels) - return backbone_features, proj_features, labels - - -@torch.no_grad() -def run_knn( - train_features: torch.Tensor, - train_targets: torch.Tensor, - test_features: torch.Tensor, - test_targets: torch.Tensor, - k: int, - T: float, - distance_fx: str, -) -> Tuple[float]: - """Runs offline knn on a train and a test dataset. - - Args: - train_features (torch.Tensor, optional): train features. - train_targets (torch.Tensor, optional): train targets. - test_features (torch.Tensor, optional): test features. - test_targets (torch.Tensor, optional): test targets. - k (int): number of neighbors. - T (float): temperature for the exponential. Only used with cosine - distance. - distance_fx (str): distance function. - - Returns: - Tuple[float]: tuple containing the the knn acc@1 and acc@5 for the model. - """ - - # build knn - knn = WeightedKNNClassifier( - k=k, - T=T, - distance_fx=distance_fx, - ) - - # add features - knn( - train_features=train_features, - train_targets=train_targets, - test_features=test_features, - test_targets=test_targets, - ) - - # compute - acc1, acc5 = knn.compute() - - # free up memory - del knn - - return acc1, acc5 - - -def main(): - args = parse_args_knn() - - # build paths - ckpt_dir = Path(args.pretrained_checkpoint_dir) - args_path = ckpt_dir / "args.json" - ckpt_path = [ - ckpt_dir / ckpt for ckpt in os.listdir(ckpt_dir) if ckpt.endswith(".ckpt") - ][0] - - # load arguments - with open(args_path) as f: - method_args = json.load(f) - cfg = OmegaConf.create(method_args) - - # build the model - model = METHODS[method_args["method"]].load_from_checkpoint( - ckpt_path, strict=False, cfg=cfg - ) - - # prepare data - _, T = prepare_transforms(args.dataset) - train_dataset, val_dataset = prepare_datasets( - args.dataset, - T_train=T, - T_val=T, - train_data_path=args.train_data_path, - val_data_path=args.val_data_path, - data_format=args.data_format, - ) - train_loader, val_loader = prepare_dataloaders( - train_dataset, - val_dataset, - batch_size=args.batch_size, - num_workers=args.num_workers, - ) - - # extract train features - train_features_bb, train_features_proj, train_targets = extract_features( - train_loader, model - ) - train_features = {"backbone": train_features_bb, "projector": train_features_proj} - - # extract test features - test_features_bb, test_features_proj, test_targets = extract_features( - val_loader, model - ) - test_features = {"backbone": test_features_bb, "projector": test_features_proj} - - # run k-nn for all possible combinations of parameters - for feat_type in args.feature_type: - print(f"\n### {feat_type.upper()} ###") - for k in args.k: - for distance_fx in args.distance_function: - temperatures = args.temperature if distance_fx == "cosine" else [None] - for T in temperatures: - print("---") - print( - f"Running k-NN with params: distance_fx={distance_fx}, k={k}, T={T}..." - ) - acc1, acc5 = run_knn( - train_features=train_features[feat_type], - train_targets=train_targets, - test_features=test_features[feat_type], - test_targets=test_targets, - k=k, - T=T, - distance_fx=distance_fx, - ) - print(f"Result: acc@1={acc1}, acc@5={acc5}") - - -if __name__ == "__main__": - main() diff --git a/solo-learn/main_linear.py b/solo-learn/main_linear.py index 9de6d91..97d5f94 100644 --- a/solo-learn/main_linear.py +++ b/solo-learn/main_linear.py @@ -64,11 +64,14 @@ def main(cfg: DictConfig): # remove fc layer backbone.fc = nn.Identity() cifar = cfg.data.dataset in ["cifar10", "cifar100"] - if cifar: - backbone.conv1 = nn.Conv2d( - 3, 64, kernel_size=3, stride=1, padding=2, bias=False - ) - backbone.maxpool = nn.Identity() + + # These lines was present in the original code, but it gave an error + + # if cifar: + # backbone.conv1 = nn.Conv2d( + # 3, 64, kernel_size=3, stride=1, padding=2, bias=False + # ) + # backbone.maxpool = nn.Identity() ckpt_path = cfg.pretrained_feature_extractor assert ( diff --git a/solo-learn/main_umap.py b/solo-learn/main_umap.py deleted file mode 100644 index 0477178..0000000 --- a/solo-learn/main_umap.py +++ /dev/null @@ -1,75 +0,0 @@ -# Copyright 2023 solo-learn development team. - -# Permission is hereby granted, free of charge, to any person obtaining a copy of -# this software and associated documentation files (the "Software"), to deal in -# the Software without restriction, including without limitation the rights to use, -# copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the -# Software, and to permit persons to whom the Software is furnished to do so, -# subject to the following conditions: - -# The above copyright notice and this permission notice shall be included in all copies -# or substantial portions of the Software. - -# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, -# INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR -# PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE -# FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR -# OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER -# DEALINGS IN THE SOFTWARE. - -import json -import os -from pathlib import Path - -from omegaconf import OmegaConf - -from solo.args.umap import parse_args_umap -from solo.data.classification_dataloader import prepare_data -from solo.methods import METHODS -from solo.utils.auto_umap import OfflineUMAP - - -def main(): - args = parse_args_umap() - - # build paths - ckpt_dir = Path(args.pretrained_checkpoint_dir) - args_path = ckpt_dir / "args.json" - ckpt_path = [ - ckpt_dir / ckpt for ckpt in os.listdir(ckpt_dir) if ckpt.endswith(".ckpt") - ][0] - - # load arguments - with open(args_path) as f: - method_args = json.load(f) - cfg = OmegaConf.create(method_args) - - # build the model - model = ( - METHODS[method_args["method"]] - .load_from_checkpoint(ckpt_path, strict=False, cfg=cfg) - .backbone - ) - # prepare data - train_loader, val_loader = prepare_data( - args.dataset, - train_data_path=args.train_data_path, - val_data_path=args.val_data_path, - data_format=args.data_format, - batch_size=args.batch_size, - num_workers=args.num_workers, - auto_augment=False, - ) - - umap = OfflineUMAP() - - # move model to the gpu - device = "cuda:0" - model = model.to(device) - - umap.plot(device, model, train_loader, "im100_train_umap.pdf") - umap.plot(device, model, val_loader, "im100_val_umap.pdf") - - -if __name__ == "__main__": - main() diff --git a/solo-learn/zoo/cifar10.sh b/solo-learn/zoo/cifar10.sh deleted file mode 100644 index 9fc623f..0000000 --- a/solo-learn/zoo/cifar10.sh +++ /dev/null @@ -1,109 +0,0 @@ -mkdir trained_models -cd trained_models -mkdir cifar10 -cd cifar10 - -# Barlow Twins -mkdir barlow_twins -cd barlow_twins -gdown https://drive.google.com/uc?id=1x7y44E05vuobibfObT4n3jqLI8QNVESV -gdown https://drive.google.com/uc?id=1Mxfq2YGQ53bNRV2fNYzvYIneM5ZGeb2h -cd .. - -# BYOL -mkdir byol -cd byol -gdown https://drive.google.com/uc?id=1zOE8O2yPyhE23LMoesMoDPdLyh1qbI8k -gdown https://drive.google.com/uc?id=1l1XIWE1ailKzsQnUPGDgyvK0escOsta6 -cd .. - -# DeepCluster V2 -mkdir deepclusterv2 -cd deepclusterv2 -gdown https://drive.google.com/uc?id=13L_QlwrBRJhdeCaVdgkRYWfvoh4PIWwj -gdown https://drive.google.com/uc?id=17jRJ-LC56uWRuNluWXecXHjTxomuGs_T -cd .. - -# DINO -mkdir dino -cd dino -gdown https://drive.google.com/uc?id=1Wv9w5j22YitGAWi4p3IJYzLVo4fQkpSu -gdown https://drive.google.com/uc?id=1PBElgMN5gjZsK3o1L55jNnb5A1ebbOvu -cd .. - -# MoCo V2+ -mkdir mocov2plus -cd mocov2plus -gdown https://drive.google.com/uc?id=1viIUTHmLdozDWtzMicV4oOyC50iL2QDU -gdown https://drive.google.com/uc?id=1ZLpgK13N8rgBxvqRbyGFd_8mF03pStIx -cd .. - -# MoCo V3 -mkdir mocov3 -cd mocov3 -gdown https://drive.google.com/uc?id=1EFHWBLYFsglZYPYsBc0YrtihrzBZRe7h -gdown https://drive.google.com/uc?id=1Gb6TCWoY2aN8AK3UnuZu4IxIktqbDybP -cd .. - -# NNCLR -mkdir nnclr -cd nnclr -gdown https://drive.google.com/uc?id=1zKReUmJ35vRnQxfSxn7yRVRW_oy3LUDF -gdown https://drive.google.com/uc?id=1UyI9r19PoFGqHjd5r1UEpstCSTkleja7 -cd .. - -# ReSSL -mkdir ressl -cd ressl -gdown https://drive.google.com/uc?id=1UdDWvgpyvj3VFVm0lq-WrGj0-GTcEpHq -gdown https://drive.google.com/uc?id=1XkkYUuEI79__4GpCCDhuFEbv0BbRdCBh -cd .. - -# SimCLR -mkdir simclr -cd simclr -gdown https://drive.google.com/uc?id=15fI7gb9M92jZWBZoGLvarYDiNYK3RN2O -gdown https://drive.google.com/uc?id=1HMJof4v2B5S-khepI_x8bgFv72I5KMc9 -cd .. - -# Simsiam -mkdir simsiam -cd simsiam -gdown https://drive.google.com/uc?id=1ZMGGTziK0DbCP43fDx2rPFrtJxCLJDmb -gdown https://drive.google.com/uc?id=1hh1QrQiWfRej-8D6L67T_F7Je9-EUUg2 -cd .. - -# SupCon -mkdir supcon -cd supcon -gdown https://drive.google.com/uc?id=1tkk_r7tYozLgf9khW6LiGxaTvJQ4c5sA -gdown https://drive.google.com/uc?id=1OhZul-rtBVUOqvOkORk8HgOLIXGCNEzB -cd .. - -# SwAV -mkdir swav -cd swav -gdown https://drive.google.com/uc?id=1CPok55wwN_4QecEjubdLeBo_9qWSJTHw -gdown https://drive.google.com/uc?id=1t59f1Q8ifx8tAySGpD2pmvogNcR1USEo -cd .. - -# VIbCReg -mkdir vibcreg -cd vibcreg -gdown https://drive.google.com/uc?id=1dHsKrhCcwWIXFwQJ4oVPgLcEcT3SecQV -gdown https://drive.google.com/uc?id=1OPsUf8VnKo5w6T8-rEQFaodUNxvQ8CTT -cd .. - -# VICReg -mkdir vicreg -cd vicreg -gdown https://drive.google.com/uc?id=1TeliMNt5bOchqJj2u_JjB0_ahKB5LKi5 -gdown https://drive.google.com/uc?id=1dsdPL-5QNS9LyHypYN6VQfEuiNWLKJqN -cd .. - -# W-MSE -mkdir wmse -cd wmse -gdown https://drive.google.com/uc?id=1jTjpmVTi9rtzy3NPEEp_61py-jeHy5fi -gdown https://drive.google.com/uc?id=1YLuqazfSDOruSiu4Kl6OAexDnt5LKEIT -cd .. diff --git a/solo-learn/zoo/cifar100.sh b/solo-learn/zoo/cifar100.sh deleted file mode 100644 index 35673bb..0000000 --- a/solo-learn/zoo/cifar100.sh +++ /dev/null @@ -1,109 +0,0 @@ -mkdir trained_models -cd trained_models -mkdir cifar100 -cd cifar100 - -# Barlow Twins -mkdir barlow_twins -cd barlow_twins -gdown https://drive.google.com/uc?id=17cZt3DorfiCYb0ZauLHv0iM-YDGYa-mE -gdown https://drive.google.com/uc?id=17Me99dh-XfTV-fniXn0Cy-ZcwGa9dRZe -cd .. - -# BYOL -mkdir byol -cd byol -gdown https://drive.google.com/uc?id=1fE7TdRboFJnYXr8JSY_tGmuFGitI8l23 -gdown https://drive.google.com/uc?id=1qsBJoO1ROAEUeQtvl8hOBDnLXZKY8Ziy -cd .. - -# DeepCluster V2 -mkdir deepclusterv2 -cd deepclusterv2 -gdown https://drive.google.com/uc?id=1grFfh0aaVYpeuYbgFYB4rmfj9uvXhYSd -gdown https://drive.google.com/uc?id=12jBsv8Fd2vk6OD5khbl4qp7szfSCiERD -cd .. - -# DINO -mkdir dino -cd dino -gdown https://drive.google.com/uc?id=16gdp5L_a9BVcRvcU4f-NUJCsIpX3Oecr -gdown https://drive.google.com/uc?id=1M4UVug_ARfNW_sjnRbc0KBceBXVKVVxH -cd .. - -# MoCo V2+ -mkdir mocov2plus -cd mocov2plus -gdown https://drive.google.com/uc?id=1KNkCA2Hr70QsmOSif9_UUndFerOb7Jft -gdown https://drive.google.com/uc?id=1T_SpFAEhZap2fvKnUvk8hzL-C7Nzad93 -cd .. - -# MoCo V3 -mkdir mocov3 -cd mocov3 -gdown https://drive.google.com/uc?id=1QAuKJmegGCJrntAL80tfTrbi2fI4sPl- -gdown https://drive.google.com/uc?id=1jtJEi66g5z7dBn0FDcSL7zoU4ArEEyqU -cd .. - -# NNCLR -mkdir nnclr -cd nnclr -gdown https://drive.google.com/uc?id=1aodwBlGK6EqrC_kthk8JcuxVcY4S5CF9 -gdown https://drive.google.com/uc?id=14Z8REvCrdW8eZ0kwxmNioIPneyCSAk0E -cd .. - -# ReSSL -mkdir ressl -cd ressl -gdown https://drive.google.com/uc?id=16sKNdpScv5FckpC02W41mjETXL6T5u2S -gdown https://drive.google.com/uc?id=1niA588wO6KX1dcbhfelb_vumByHgDfVV -cd .. - -# SimCLR -mkdir simclr -cd simclr -gdown https://drive.google.com/uc?id=17YGC7y4yxkVAF8ZNezdtmN-uc70jz3zq -gdown https://drive.google.com/uc?id=1bmrfJxEK505_ky0m7q7ZJSDpFfgqIuQ6 -cd .. - -# Simsiam -mkdir simsiam -cd simsiam -gdown https://drive.google.com/uc?id=1DStn9PAEMJtzh1Mxb3NjfTtm5vaNgRM5 -gdown https://drive.google.com/uc?id=1y03EtFuMi5fZGPJZfN3hkONe99WsFBOJ -cd .. - -# SupCon -mkdir supcon -cd supcon -gdown https://drive.google.com/uc?id=1QhPHENtgYttIF1Dn1srA4dAkIiC_5P7W -gdown https://drive.google.com/uc?id=1QsZs9TfWoycrHBBUrliWe-cqkGQ9epAD -cd .. - -# SwAV -mkdir swav -cd swav -gdown https://drive.google.com/uc?id=1oJzFfayNpcShK1bZtDK58HthcKY2bpns -gdown https://drive.google.com/uc?id=14ed_7MG_pg-G_qjQcxVc8MUZWwFcz3mF -cd .. - -# VIbCReg -mkdir vibcreg -cd vibcreg -gdown https://drive.google.com/uc?id=1akNcewHzh4ideoQPWakaXWGDxfGoxkNu -gdown https://drive.google.com/uc?id=1cdvZXUmmDptSe-RkYyiQXyREwthvMuxW -cd .. - -# VICReg -mkdir vicreg -cd vicreg -gdown https://drive.google.com/uc?id=1kH78BUBKprrsxL2KRKmorVQ9vJHsMsID -gdown https://drive.google.com/uc?id=1TJk6G6KY1URPpruhKIDuovv66U-mnQHo -cd .. - -# W-MSE -mkdir wmse -cd wmse -gdown https://drive.google.com/uc?id=1_6EmYFqAW_U8DFv72KUaAe-BV8xkRxsp -gdown https://drive.google.com/uc?id=1uIeg5EKEMefeBIyYFm9SBmChJPBc-0g_ -cd .. diff --git a/solo-learn/zoo/imagenet.sh b/solo-learn/zoo/imagenet.sh deleted file mode 100644 index a875703..0000000 --- a/solo-learn/zoo/imagenet.sh +++ /dev/null @@ -1,43 +0,0 @@ -mkdir trained_models -cd trained_models -mkdir imagenet -cd imagenet - -# Barlow Twins -mkdir barlow_twins -cd barlow_twins -gdown https://drive.google.com/uc?id=1GodHwmdMn9u76b5XFzEr5v59tOfwUOof -gdown https://drive.google.com/uc?id=1EKdbR72-gtNE782254tjXi9UR2NiwEWh -cd .. - -# BYOL -mkdir byol -cd byol -gdown https://drive.google.com/uc?id=1TheL_4tmDWByCxg8XHke5VEz_lcYHH64 -gdown https://drive.google.com/uc?id=18gG0Jo59cFVX4qNUO119jIhzHcJkAmGz -cd .. - -# MoCo V2+ -mkdir mocov2plus -cd mocov2plus -gdown https://drive.google.com/uc?id=1BBauwWTJV38BCf56KtOK9TJWLyjNH-mP -gdown https://drive.google.com/uc?id=1JMpGSYjefFzxT5GTbEc_2d4THxOxC3Ca -cd .. - -# MAE -mkdir mae -cd mae - -mkdir pretrain -cd pretrain -gdown https://drive.google.com/uc?id=1WfkMVNGrQB-NK12XPkcWWxJsFy1H_0TI -gdown https://drive.google.com/uc?id=1EAeZy3lyr35wVcPBISKQXjHFxXtxA0DY -cd .. - -mkdir finetune -cd finetune -gdown https://drive.google.com/uc?id=1buWWhf7zPJtpL3qOG_LRePfnwurjoJtw -gdown https://drive.google.com/uc?id=1n6symLssKGolf_WQd5I1RS-Gj5e-go92 -cd .. - -cd .. diff --git a/solo-learn/zoo/imagenet100.sh b/solo-learn/zoo/imagenet100.sh deleted file mode 100644 index 44d632f..0000000 --- a/solo-learn/zoo/imagenet100.sh +++ /dev/null @@ -1,123 +0,0 @@ -mkdir trained_models -cd trained_models -mkdir imagenet100 -cd imagenet100 - -# Barlow Twins -mkdir barlow_twins -cd barlow_twins -gdown https://drive.google.com/uc?id=1C2qQSqp8cXvfrwHVG9MuGTPT2TOTsGla # checkpoint -gdown https://drive.google.com/uc?id=1TY10aa97P4Fl7EgSjTy_u_QME9tkcU4r # args -cd .. - -# BYOL -mkdir byol -cd byol -gdown https://drive.google.com/uc?id=1cgJaSRr3HPZRNMwzYwwS5Vwtkna3LgGs # checkpoint -gdown https://drive.google.com/uc?id=1EIluSRGaV0Ft1UQecGhpkFUCKVwMmtv9 # args -cd .. - -# DeepCluster V2 -mkdir deepclusterv2 -cd deepclusterv2 -gdown https://drive.google.com/uc?id=1ANWOVMFMa-9eRWTKRGiUkNenJYD-McjT # checkpoint -gdown https://drive.google.com/uc?id=18oOypleOOHQ7z9XL9zUTgDB7zpRdbMti # args -cd .. - -# DINO -mkdir dino -cd dino -gdown https://drive.google.com/uc?id=1MkuNjlIMqzuRwdG_K6NoDrGQH2GtssXV # checkpoint -gdown https://drive.google.com/uc?id=1MlYaqPsp_pEaDR7nDRbxv3oOsMTVBHg9 # args -cd .. - -# DINO (vit tiny) -mkdir dino-vit -cd dino-vit -gdown https://drive.google.com/uc?id=11rHOKD4EQB2AJ1C2tLHz0pjai6MqwT9v # checkpoint -gdown https://drive.google.com/uc?id=15pQbMd0xiLZNsmozBmsKA3_HVEdpSqmy # args -cd .. - -# MoCo V2+ -mkdir mocov2plus -cd mocov2plus -gdown https://drive.google.com/uc?id=1aXGypKbIqV8BqtVOzpk2lRJWXb--XejO # checkpoint -gdown https://drive.google.com/uc?id=1s5rzHSqAMRKaUR4ZP3HWbCm8QLXU6JQ8 # args -cd .. - -# MoCo V3 -mkdir mocov3 -cd mocov3 -gdown https://drive.google.com/uc?id=1cUaAdx-6NXCkeSMo-mQtpPnYk7zA4Gg4 # checkpoint -gdown https://drive.google.com/uc?id=1mb6ZRKF1CdGP0rdJI2yjyStZ-FCFjsi4 # args -cd .. - -# MoCo V3 R50 -mkdir mocov3-r50 -cd mocov3-r50 -gdown https://drive.google.com/uc?id=1KiwHisYRmzYjLYDm1zQxZlUKe8BkI2i8 # checkpoint -gdown https://drive.google.com/uc?id=16pix6gNybXnssMpXlzjKnMWl9lRfdv20 # args -cd .. - -# NNCLR -mkdir nnclr -cd nnclr -gdown https://drive.google.com/uc?id=1rj9-YBUNX0wHVLjQuksOOubfEZJsrsjF # checkpoint -gdown https://drive.google.com/uc?id=1GBT6-QkhuDLexfVgwM0SWbzuXJ6QrF9o # args -cd .. - -# ReSSL -mkdir ressl -cd ressl -gdown https://drive.google.com/uc?id=1AH3hFcakrGKXzxmzO2LBHWjk-Mgu5PUN # checkpoint -gdown https://drive.google.com/uc?id=1XWKERLv_YgFQ_Oy33TD8DhTfH9qSoVQa # args -cd .. - -# SimCLR -mkdir simclr -cd simclr -gdown https://drive.google.com/uc?id=1dU88Sh5F_8J_UXXEQ8FOWS85g8eFEZVa # checkpoint -gdown https://drive.google.com/uc?id=1865vcQhuvGeNm0iQ9g87APLwLvYNuqcn # args -cd .. - -# Simsiam -mkdir simsiam -cd simsiam -gdown https://drive.google.com/uc?id=1cwAyDCpU36zmQ6-r4Ww7YqiZ41vbjckQ # checkpoint -gdown https://drive.google.com/uc?id=1EU43HZKrLu_ZTV3CVAkjkFS6HORhtmR9 # args -cd .. - -# SupCon -mkdir supcon -cd supcon -gdown https://drive.google.com/uc?id=1-NRvw7J9WrQKBvDhmuirQmTklMlQasxI # checkpoint -gdown https://drive.google.com/uc?id=1IKTW20UTWlHSO4RsgO1QxakYs5ZscY26 # args -cd .. - -# SwAV -mkdir swav -cd swav -gdown https://drive.google.com/uc?id=1nDiXHb8ce6_qDyZ8EcqDXi6ptI4A_t6B # checkpoint -gdown https://drive.google.com/uc?id=1h1-YEqEw5Zj7wl0Gkxiz6WC3qpwa2FgL # args -cd .. - -# VIbCReg -mkdir vibcreg -cd vibcreg -gdown https://drive.google.com/uc?id=1VDUvp0zghvnUgwhWS-s7PuCA1KTPEPPX # checkpoint -gdown https://drive.google.com/uc?id=14rEyW3cZyUxctjLQunIjyuJMI3DbUQ-b # args -cd .. - -# VICReg -mkdir vicreg -cd vicreg -gdown https://drive.google.com/uc?id=1yAxL-NTOYN6kGi2VtPeo7cPKFxXFKYyP # checkpoint -gdown https://drive.google.com/uc?id=1A5QaOlUGaId3qECmQusoPDiYx8tjfKDz # args -cd .. - -# W-MSE -mkdir wmse -cd wmse -gdown https://drive.google.com/uc?id=1yYhOsIpbHqJGhqlbMTYwBxMOJkz7rSwo # checkpoint -gdown https://drive.google.com/uc?id=1Q88g4Rtz_k4FR9QvXwFqAZ-kYuL4dXl- # args -cd .. From b27c52bbbfaff7e38f2095991fa4d6ac57d62ab9 Mon Sep 17 00:00:00 2001 From: sanaAyrml Date: Thu, 16 Jan 2025 07:18:20 -0500 Subject: [PATCH 6/7] Update readme --- .pre-commit-config.yaml | 43 --------------------------- CONTRIBUTING.md | 31 -------------------- README.md | 46 +++++++++++++++++------------ images/GenSSL_last-main.png | Bin 0 -> 1460900 bytes pyproject.toml | 56 ------------------------------------ requirements.txt | 10 ------- 6 files changed, 27 insertions(+), 159 deletions(-) delete mode 100644 .pre-commit-config.yaml delete mode 100644 CONTRIBUTING.md create mode 100644 images/GenSSL_last-main.png delete mode 100644 pyproject.toml delete mode 100644 requirements.txt diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml deleted file mode 100644 index 51de3a9..0000000 --- a/.pre-commit-config.yaml +++ /dev/null @@ -1,43 +0,0 @@ -repos: - - repo: https://github.com/pre-commit/pre-commit-hooks - rev: v4.4.0 # Use the ref you want to point at - hooks: - - id: trailing-whitespace - - id: check-ast - - id: check-builtin-literals - - id: check-docstring-first - - id: check-executables-have-shebangs - - id: debug-statements - - id: end-of-file-fixer - - id: mixed-line-ending - args: [--fix=lf] - - id: requirements-txt-fixer - - id: check-yaml - - id: check-toml - - - repo: https://github.com/astral-sh/ruff-pre-commit - rev: v0.1.5 - hooks: - - id: ruff - args: [--fix] - types_or: [python, jupyter] - - - repo: https://github.com/psf/black - rev: 23.7.0 - hooks: - - id: black - - - repo: https://github.com/pre-commit/mirrors-mypy - rev: v1.6.1 - hooks: - - id: mypy - entry: python -m mypy --show-error-codes --pretty --config-file pyproject.toml - - - repo: https://github.com/nbQA-dev/nbQA - rev: 1.7.0 - hooks: - - id: nbqa-black - - id: nbqa-ruff - args: [--fix] - -exclude: 'icgan/.*|rcdm/.*' diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md deleted file mode 100644 index e67db36..0000000 --- a/CONTRIBUTING.md +++ /dev/null @@ -1,31 +0,0 @@ -# Contributing to AI Engineering Projects - -Thanks for your interest in contributing! - -To submit PRs, please fill out the PR template along with the PR. If the PR -fixes an issue, don't forget to link the PR to the issue! - -## Pre-commit hooks - -Once the python virtual environment is setup, you can run pre-commit hooks using: - -```bash -pre-commit run --all-files -``` - -## Coding guidelines - -For code style, we recommend the [google style guide](https://google.github.io/styleguide/pyguide.html). - -Pre-commit hooks apply the [black](https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html) -code formatting. - -For docstrings we use [numpy format](https://numpydoc.readthedocs.io/en/latest/format.html). - -We also use [flake8](https://flake8.pycqa.org/en/latest/) and [pylint](https://pylint.pycqa.org/en/stable/) -for further static code analysis. The pre-commit hooks show errors which you need -to fix before submitting a PR. - -Last but not the least, we use type hints in our code which is then checked using -[mypy](https://mypy.readthedocs.io/en/stable/). Currently, mypy checks are not -strict, but will be enforced more as the API code becomes more stable. \ No newline at end of file diff --git a/README.md b/README.md index 7fe5fd3..f581f28 100644 --- a/README.md +++ b/README.md @@ -1,20 +1,29 @@ # Generative SSL -This is the PyTorch implemention of our paper **"Can Generative Models Improve Self-Supervised Representation Learning?"** submitted to ECCV 2024 for reproducing the experiments. +This repository contains the PyTorch implementation of **"Can Generative Models Improve Self-Supervised Representation Learning?"** accepted to AAAI 2025. + +## Abstract + +Self-supervised learning (SSL) holds significant promise in leveraging unlabeled data for learning robust visual representations. However, the limited diversity and quality of existing augmentation techniques constrain SSL performance. We introduce a novel framework that incorporates generative models to produce semantically consistent and diverse augmentations conditioned on source images. This approach enriches SSL training, improving downstream task performance by up to 10\% in Top-1 accuracy across various techniques. + +![avatar](./images/GenSSL_last-main.png) +Our augmentation pipeline utilizes generative models, i.e., Stable Diffusion or ICGAN, conditioned on the source image representation, accompanied by the standard SSL augmentations. The components inside the Generative Augmentation module, i.e. the pretrained SSL encoder and the generative model remain frozen throughout the SSL training process. + ## Requirements -To create the virtual environment for running the experiments, you need to run: +We used solo-learn library for the implementation of SSL method. You can find the library in this [LINK](https://github.com/vturrisi/solo-learn). -`pip install -r requirements.txt` +To create the virtual environment for running the experiments please first: -**Note:** -**You always need to set the proper path to the virtual environment, the dataset and the model in each SLURM file before submitting the job. Here are the options for the datasets and models that we used in our experiments:** +`cd solo-learn` + +Then install requirements based on solo-learn library documentation [here](https://github.com/vturrisi/solo-learn?tab=readme-ov-file#installation). -- **Datasets:** ImageNet, iNaturalist2018, Food101, Places365, CIFAR10/100 -- **Models:** Baseline (SimSiam model trained on ImageNet), SimSiam model trained with ICGAN augmentations, SimSiam model trained with Stable Diffusion augmentations ## Data Generation +**Note:** +**You always need to set the proper path to the virtual environment and path to save generated data in generation scripts.** To generate augmentations with ICGAN run: @@ -24,24 +33,23 @@ To generate augmentations with Stable Diffusion run: `sbatch GenerativeSSL/scripts/generation_scripts/gen_img_stablediff.slrm` -## Training - -To train the SimSiam method on the ImageNet, run: - -`sbatch GenerativeSSL/scripts/train_scrpits/train_simsiam_singlenode.slrm` +## Training and Evaluation +**Note:** +**You always need to set the proper path to the virtual environment in solo-learn slrm files. We pretrained our models on train split of Imagenet. Here is the model and dataset choices for the evaluation that we used in our experiments:** -In this file, there is a `use_synthetic_data` flag that you can use to train the model with augmentations. You just need to specify the path to synthetic data. (Either ICGAN or Stable Diffusion augmentations) By default, the `use_synthetic_data` flag has been passed in the SLURM file. +- **Datasets:** ImageNet, iNaturalist2018, Food101, Places365, CIFAR10/100 +- **Models:** SimCLR (Baseline, ICGAN, Stablediff), SimSiam (Baseline, ICGAN, Stablediff), MoCo (Baseline, ICGAN, Stablediff), BYOL (Baseline, ICGAN, Stablediff), Barlow Twins (Baseline, ICGAN, Stablediff) -## Evaluation +### Training -For downstream tasks, there are all evaluation scripts in this `GenerativeSSL/scripts/eval_scripts` folder. In each dataset folder in `eval_scripts` there are three SLURM files. (baseline model, model trained with ICGAN aug, model trained with stablediff aug) +Configs for training are in the `solo-learn/scripts/pretrain` folder. You can find the config files for each model and dataset in the respective folders. You need to set **path for the dataset** and **dir to save model** in each respective config file before submitting the job. By choosing the desired config you can train the methods on the ImageNet, run: -Similarly for evaluation, you just need to submit the slurm file related to the dataset you want. Again, you need to specify the path to the virtual environment, the dataset and the related checkpoint in each SLURM file. For example, command below run the experiment of evaluating model trained with stable diffusion augmentations on Food101: +`sbatch scripts/solo_learn/train_solo_learn.slrm` -`sbatch GenerativeSSL/scripts/eval_scripts/food101/stablediff.slrm` -## Pretrained Models -We also provide the checkpoints for all the trained models here in the [LINK](https://drive.google.com/drive/folders/1xPIbf1cOPqzIzuZ185GjAprA8XmQ0Tvu) +### Evaluation +Configs for evaluation are in the `solo-learn/scripts/linear` folder. You can find the config files for each model and dataset in the respective folders. You need to set **path for the dataset**, **dir to save model** and **path to pretrained feature extractor** in each respective config file before submitting the job. By choosing the desired config you can train the methods on the ImageNet, run: +`sbatch scripts/solo_learn/eval_solo_learn.slrm` diff --git a/images/GenSSL_last-main.png b/images/GenSSL_last-main.png new file mode 100644 index 0000000000000000000000000000000000000000..67c96919b5e6ee2b1ac22b26c565f08333ca29d7 GIT binary patch literal 1460900 zcmce;xvuow)-JSv4nhcthzyZ#368{;o3dRF-S>96w<|ev+2y9(_stC;ZUUVvAm9eb z2uVoE5jhvY@b`WD=l~&zS8Kg%EjQECvl-(V&v-`tw=YxrPydtu{6G2I-~RTWDl+f< z?Qj3;fBM_s{zdqY|06i^kAMAt{CDue`zFX)h+#8{vBuGZ>3wl4g7?|@W184e-V76$iG2}e?$FO z@GCP6%{Tmhv6_O5eN*tA=Bn@D2K4!l zyGw48CsXx*Y2ZG*{dL_RFN2Z)K>k7g`3eKP!~Ewv!n5v6=QL@)e;z^qJo4$M>^uE= z2;L3*J1+bkVc^pH3+MhC7x<$~+Q05>lL|U@H&z^&QhcqQaY4q72W0oC%Hh~fN^Up@J^REsP#9v2>>7Nh!tJC)Zx4QoMCV%uBLyEP^ z{`tOtUAXjpvsA-hgYehEwC}oa`q$k4bu5{te*eE-miNtHulLtA_i+BzwLg1$A8`L4 zBYGbVt89VAaj!0#{A=v*46g-uR4U{_DRU4$JHRqXz%ouucDN zC;k8PU|5=e?2DvX|C+Er56sK?&(#k@JlyB1_K!(rVGk`8Or@TDzRjzjtL5H#IRELF zrEg)IH}?rPN$X z=ATDli~V`zZYo)lzvIlGJ;KD_@GbtZ3nKl4Ie9RQ?bdJADHxWq1;%Aqg<)WUiSSjh zvW=>u9XK!l*B7A}7eO1GDb{SCin;WkuqzzvzN45_c;VD90n@JjMQP?7u*}_-YJ(W4TR=8sKd3)EneSZt3@e)}i zk&j0vm3yji_NtytrE>;cT;Q6!?}le1Cj0Cx4|y2BeKjuo{AcTe3N0GYQq0>(Y*a$3 zc*<$~{@;B=>raLe?1H@tb!2v-Vi?6?Y{Ox4I43ad&cJyQ#*~F${=I)b%M^yd7L3fj z{{5-!BSM57y~p=5_bdW~5zU;F*8S@4wfP`xwl=EwaVk<%3}m!DhHWo?kE` zQ}{f~z+Lx08{X{{T*G!))-ftS6wcIfp)b;5FVLSxXP=J$dwBONVFhErh%ihk@X#V0 zVP0&AA~hgPClGvoJ~Ww#|$4md`UArv0P`FDq$wd;f-mFp)i`2rP^-lv&sHC zRu_YvAtWV;+o+D#D~X9RBZE)Q;XI4KshEmfO7#5WtqV#VWX}}EMy%f_?uV({_~K=8 zv@(;eL`>Q7A(&It`pU*xuq}(kfrSafYV-D^AFp|8%KS>u(43TKKUn-RW^M)?K@yT zUudULH5?Nhxl!I8s1>rX z(16#TXC?G3b6KHW5T9f@2~{GvQV6_(^5@3U-yf++nioeKVJv8|!(L9n`d=lw&5gmv zFa%lnBry)#KEfH$zxq%##UKYQcG2ebu1ArhcMTiRKC)E8TRinJW)_h&ftk*&i9a(m zbjcUV3!3ue!uo+2ySvy5Yx%Qxp|O#O+f>GBzPwYnQsPWwde!6dO^lWHs^8Jn8PutMK`EFR<3@{2aJLY22{ zwY)pbPSIK6%Iw1{$5rjj@ql+0O91BZdHix0)d_yrsY-7804NGgJ zH!S^Avp&CA>Oc7y^?&riI~t8^jle5D zjkLzh+kCwy@?og00397gr=}^LKUI4y)OfpMeutKaDiE%nS*0^5>JhF>bexeiFF{VH zMn?AeWsUtw-BUX*^%x;H3~rC9tX2B@-$4A{3ZbJ7?)hh4VgK=c>En@wbFJ`k= zy=|`10_I(i)u8AG|A1j(qr!0WaV1Y*)Iz@)-F8=z=XK#}$E=(UoA~`YN2`gTX@;fx zcRgJ;$`muj>efCH*^Am=QfbVoAJ#|uM1`tRxPY12u-l?m%0tITyK)F>y}b%6dextf zAx(jMJsGJ9-!;xK>^&efOdeUYd+1SNel11>G* zSZz2QKD7#Ugwyv2tUpiygyM94ME{Oi9#2A-je6iwx&t%BPC~b;eyveuT$07+;iq)P zoiNx=8VKj4_|5U`6XV zjD>h@JV#Wg_!V}T{tdD*{n+w}s$oL2_Q1t`rah|fRsQJ9A_79baC9C+r9N9}kF{93 zR3`13QG59$DjB4TvK}9o57!u7?$VTTLTRU>bcgLK$0d2pt%*h^!0LXa7F|qa<&F9o ze@dq)-j5&aHg7Jo*<=aPXNT~t^0fE&`qIJdyQ1)R1?x=Owkx9V67`>&(}c4~4drvB zqd%vUIEMFD?W&(_Sh{XE6B~$JQP!(iJ>Id=48$>r)f# z$m1=T4EKZBpQUqd`xQf?x7Djk_f@!k^mGl+KNAhhh^AT_*FF<>3xYy@J!8cumQj9G~nSMY7 z&)6TM?;xWilq0WE7dOl(RcGGg@qwnwqSngAURe*ujIr!z@({rCmYR~99jn?l@6s;Y zV^?BMu49m+66|)bDpp`?c8K!N0?p^gVe!iI$RYW$;dCr9p=x$XGu?7`k4h7B^;y5}K zoml32PL-^1mR@5s;WgMA@jGAZV+#9ERUfKT6Ay*0YLrzK#agy}sGriwFMCu=lk#EQQdf-RI;0dN{ z`LmeXD+Ig54hz6UakwmrRdTJ@`I=$391zQC*6*Otbp_r-Ylw(>AD^A3l1EqBo>0;4 zvLYiB%C8&o7fB~3JVUvP&&-?l;F;~b{ST*AtG40X9F6{~b7#BHu-}$7QH_$? z(9=iF#xqxBR>oVuF^#DVg!rO2XvXOeTD?R8{5j^laAlO{XU{dkX$&j?c*3-jHiP#Y zt>v}|OBw9oYb-dyPznMfrfZgeeggRZ(uELI8tf!3 z^hd^(4Fq~2xnH$hqgh0MptrCax3DaJic8je*|LgZ3q^t?1x{xbY1=*+EjQA--~kuI z!#;6MnTDmn?!bM|7U7Us0w|h}N1G*kJpyYO@CY{LuO>FOl7 z+Yj}a97`#Vpwsy^WAd$PT^9{xBVoI|zV6kP3rPUFh>IQwGVP!h{6un!Eq?_4z?cJ z>LnrDGkyxXr;*uMvM|I_W|4VvL=ZgfqTzHzM#=a)6d^AvvhjU|$9}Yo>fY16TJeCT zfubyx=x2FwqBs$g6Tp1ri&=>1;`ng@)=U;>H0+sp>BMF}Rj0z8Um2!+HX5d(c`=gy z8U*?!d}b;20pzh3`fgQ3nISS4ZO!ywbegh3>92bn9h_`>)0hvmp>Yo!=lvz;6B9iER7hJVR9%e=q%yFLeL#`*#gXH?EWxF+7xyTA zfZMHZjyN!$qEgd*_vnO@)9q^bo_*VR??uM8sZwtcZs@*71WZ-gs6*YZXL`uQCw`s7 zD|Y-*=CrH+X|Bu6jjBNO9kE1~a~>UN7xz2@(Y?dWz)3X}!zXm)-NrNyp`2Iak?BU! z=+$NRFShi0RrCx^7agpASy4YKS_fWbBvZbYEehu&-T)SLr?Sn>KtK#yPu8Q0+# zIR}4t3q&<(^7jN9n&*W>+#kKGxW*6ho45YW*X{Oc5#TxXcZix4c&y?wWU){79Lmk+ z00?7=0CmEmk8Sw<@I~EE5FHV%;WH*LWH@TNvj|aR+{r*~0_o8^q9U;5v(!6raE|kf z$CO!CijteDoG(^zv$mSMxldKh-E}_>hVhdZk0om+56fW5q*Zf>DnyJe6RD#1OMRMj z9i=o3wS+ARWKi8ok;2aeF6!u(e`j@8iWd5pwXV)`QEFZpjv?kzRqrNIdOxtH<~^C= zR-*XsIyvW0{p_UpN75PA_vP&m3JZb2Pz1W-*cTTp6uWTKYntre^?Zh<^bLhX82Zr( zq;S)=`!$Qy@2iKOp=Mo~n}^c0Aym&~^P=#c8GhGzH4(yV6Ec#VVgl}q}%MyVKTa+l~gW68Tyybx67tP^{~s8Mo-Ty7sx_qOeS=8)c$dEczr^l)>unE`e(HVzwpW%dsWs3v32Y2yod^Hi^c#@=;s>X^%&MF z+t+LFlj)X?@jU970}R3{%&v`f-Q&^4Z;>tS`B*;q_o8sXsDh@izGZ5-9C_zq@?~Sc zmG8kOowl_vB95oN?kL6IFA*>!R}Bus+1mQ~id#0FqCI@0Rx#Y0VCgEOwxlzuu|EdN z$_!;yx%v?cFKyKFE*una$zh8>n;D|4z!63F*ZK+3h!@-;#&o-(%t+7~AH`N<>YX9q zQNi}bbVQS$Rwd3m%T-_6OOw1#cjw^-EulBSo*bXzh-ZbB5`tQj3!|yrHMBhJ=9;0* zBp52)J+y9#(gch@YtEqu$j|i6clfi$K_A&6=Ld&bsrx=3v1RZcblWhlzd+24ri^&h^u?oK-3W>YE z(sv@jX&_b|4{!_{4q0=WF9-F?pJe!hSrAhAknw4T3x8Sp467>*v>5(+qh5WfQl#yJ zgJFa?vxn&WG#v)tUcZCF5l6`fDT*ABA+RpjqqjDgr>(NMynmHr<;LaM44dH^1q+?M zBCuT~w%1BN420{CDjP~<(Cg?k7amP#kD7&9vD3qf5E(Tx-P^Ih#(ldc_pIWc(?*t8 zRAqX%`0!~S^%m~Ussu)4a(>X%wRxjVG)Ov}__Dk5^MLPG)mRS3-y5e<(wdztV77Ii z&JUJJQmeRUt@_K(MQkd3NfbwwL#QjKsLe?9e?(0bjt}MvZx-sidGj8!G4<_WcS>sf zafCu>cAde8+n>OrH5Ztd*sTUFXw2fnG}!BfFFu?s&rvz~*3h-aiks>n1dPrz%2<6* zSN$Cf03)HqYs_zGCfR=Z6VrNdFW#3>dDPbzQ2A&p$wON)H5f3_|I(KC7$~-;=EW2G z#Srh90KzWrra&IA+pUo(dLXgLP{2G)FyLH8jq>kvB^yNL6(TFfA*4izA7tty?fsAY--3lxd9BmFqp?)4v0JpeR_{H|=3 zmX^I;Yh1|BUVaTUT=*4+H%oa$!3-*l?T`ExN2XHT;0C4D9JbjLw3lP_{Bn!tYm18h zdT)TV#7;=W-gs-=o-&1XZKwXpyPQWyzigC{6J{GdE@|xwxYp{qur5n=B&GdVMNQ)?@s{5$Yt5ozAg$%%-&lqw+Lx$^yvI(pU@wR&wR|i`An;-XD z@TEg}uX%0Kna0`-a1+UG18-X*>GNBHv^M~EUo}rdkPvg#98RCjAH7w^ajAHGlbyoj^h{=axIn; zqi!hPW3bF=-k`b}+39SGiH~PbfZ#F2`2}hzcJ#(>_aC+r@-Dx0ss$F2RUdRKKW5gY z0`{CCcsY|*w?4~nVuT{o&;x5erq`57Uj7cBhQ;kG{f*`v@(C%eyx~SaIMe`pepg;&W8A1CWB$Eg7Xllki5sO zVO&I)>Dz4Y6{197r+6%t&=$Y>qcrpY&cQsh>#0>{k*r;Nz>rlmAoV8mIz zI$N#O5Kb9TVybiP&H=M{*^?xuW#3BDhIiF?b%5x#={|@pnl5&}khkEK^B5=Fvj(Fg zo*ZR5$_Irx5RKGH9a|Fy)`f*^onymu4bIj5kP zbl=^yT9fcw5+lkObL5`7BaHYdnQ<7I`t#fl)ROc^Zk#D19LV5qddoLwkKc>#4q;`d z)-H(E)R^hEOiHdV`M8PvSu_Wvp8+8k;&~v98NtC0!zkY+{JtQBLgl*m{<y|PGzI`<-v8~mZBaf1chmoLC2 z1PQ9WwTv^9?eV1nL;6KgQlZp2Pe`|*v`w=B-L%n=DM&rgFY4JCVzDcFJM<3{h(>mb z<0e@tyhy4NK>!i6lKyN|&wrMEY&raUq5L@MkO{R^+h`xu=D3`W#u0Bu`2?0tWDgZ6 zp5Kfzby_`MYd0WP?>r4fP>E9bO8gsugB?K2@P#@{ksUnv443^We)ZyG7|=a>y}sKJ zeZQI z+8{rTjACgZzcd*fb~22uE{MQ8iUm_{7uZ~JXJcSblu;5}`DL5YaFLHX?2O`FIVVel zg=g_xHU{Ti{fR8TxRRqpGuz9{IpfM-+DmgGTmn)Lg`k5FUYiEok~eZN4$~jb{qrce z!`DXs6#1fPrN(dLvg(0dJ`MNk>>OBn@eia0@+0=``P0%*ozZBvtYaQ!=|*UbU~ibF zk?nRTZd!P%`eI~HHMTxogooXWCmP#$d+gp6$(p237m!q1cZee!K~=J%MQuGMFg^Eb z`|I<7JgvZiHp=U^PBe+#hrH?z8jp{p}JFxpAao0sm`D-Nf2n*PQEUY%pdymEeD2Z9k ziirEIH|e=?tCU)a0g_JnXAY(4#vOhwydAg+-3fevelKj*aKG_lJ#PE@@{o3THe$P7 zIIi}pt9Op$uN8lW?9kiDgBCN6VX*{dYx*JDBWt4w{rR=2QYmd&g16JyGf0Jiyr=KV zMZ!?Pci)(92iR-TGi`q(l#s=*u4<*Hy0tMI;0HwY2@so={qRWMN!?+dJ4aSJR!kX4 z$PA~NHA3{+8Mr=8e*}P=%DNZr)+OCbEK^MpHHH5S?{&K8571JWeC%LH9Jg^ozs?Op z#dB&2TP9*(s6ehXZ*9ky?)&C!W{=Qw_v!>#cBg}<#_0+8`0Ttc9XfN61iRr-Sc+(nFQT}2oqq*jjG*1k5{xX z!`DVWRF!r74`+;D)WaS=Ij_+t*sKk*m4e`c_~BhjigCIRUsz>^c+$_;BWEK}0SO{H zswytu3vXpl_@5RlmdPxTft5r&=x;wBuK3*`B|NpPIi7-$8ztw&l5*!C!#aoNvkf^y_&Fo1V)2(e08WYI);l&Y=-&L`H%&fCBh9ixCU`(*ZB}q{znf%i zv!{S#aq?Ki3Ngc89!6^wAiY|xD`qJ@0ZN}8^d_a>k}I83kCK%YRTV#V$V$>IZ9@sc zZ?U#3@$%5U%y0>KhHsGC#nuWn0@L2{nh}y~Y60oRoosTom;eOc;Aa$o?vxS)32poD zYRwPeR3Ycjd>HAfKcHrs7|^=jg!+2PUodf)VIpxTrdA&8T;>et-eAlGJP;>Kziwxf za$ievH-9M!iN)z-3(12ZqlA_QK;?JX1iNoUiRUwYNLPVC;6Sv#W-yRTCzo;ozq!P9 zRX}^WZd0IAI<5$GD?sjZPPmhnED4NGfMjW~2IyAtGl+Nb`90(2RBk_?IaZVFIK_g4 z z^9fs!_7qY%1^F8)PH1JpufX<<@3T2CT&`aC?%;$V0p1L9$huZA1#4s5>I};x<}n8M z5XVNjwD7q288`t~y|~aN%su;dLp8Qa?9W4L4qUq?|!$={B1k$2%E(%-Tq)^2^3@1 zm%}{IZP_Q^^$(VU*^$(0gc3;Lnx@?l68H}^+52y3EW74ALZ~dTkqN}RGm}4QE&J=I zagjpx4A6J?STM>SInCt}?gRdcL8MUIDbN&v4^Ys+R`e*arxls4`2E&D%l1rL7^swe zRE*?qErb7DDiO@`DT&zIH6~RBKbbK}xDnty9~%6#m?c!{j3)aYfxLzu4x*Eyq>!&LugG_KeFDez>$0^LAU%)|##g*D zc5{KG#hlb3y3wF|q?YU5=J-^FO4Gca$C?u#oKBsxvN=%5@M%1M+ZLPyt(@Vv%)4?V zM^(v+Loww8*Z~i$CsjaT^-PWEdCn+_->`6b+CvV}C9nXzP51();2yD6p(=!J0kSAG zr{oJXIwg9l@CjJ1gRMZuzYMp8>yZo#uKVf z=mIzgK}--d7EmR*Z4V)XG9l1w%?G`$s#gW^+`I{1=PkSqOIo;)p`K#k{vlrW@z0Y& z%|O5wpu))JhPc@iB_X8p#9^)@t(tTt;C^p=Xj37@Zd69kE@&>pb!0?e&g2&yqCQC| zl%NwRR6D5y`QEsBo7#3)UXo4xaF8JOIOt7{C@V4Gj1Hladfs6045IVLb1_2&&q_$O zs|>8Wi$LXYvoUlRChoM;W90aJ@*<$FZLs)fPpYDO-GdTszg-RurE)bf~Sm_>l06()&mhzt9N8IAX1@{AP-Vok%Rf)|LU3O=Pye9I7BI8{D zDC(x7VEf#?;eUvAsA-5vZw>)#smM)=o6&EW2OQ zd&KYg7)jRiO*O~a4->J>k`SPKXu&34?H@vXbc0?=+Z;RGF*2kVRQqSNa%`(IE1Gdv zr`Ya!!*;RkBwjvi#93O;AK9FK%sGS?ENlu^8pck4tH}GgK^mUI52G%G@Qo7~x#|gZ zI}ht;R{^9PAlhYd7z(yqt}0|N+|#N={j7+h*A~7}4$S~X5Ef>HPG45SAgv64( zeYnZNQcv+3zZsuT>ny=R5iB1ujMZ%enz))LyG=X}?R-z_J!vS2_|F?H6jh}FOVpYS zlNxF|>eZ$TBj^@bpCSdwn!!M^S7m^*%q)HOSM5C>Ufl@Ht{n>qI8%XvD||!@(;pOP ze0!-RfamoIm!UaXmmUXq+=kEJ9`RS(=A`N30Hp!^D&Mh^l+i;a*BCP@ zFf+r=q?#b*9z!;Aw_Q{g^K5oC7bt#?0K%Q5jTaPxIJ{O3W9=vCLCuD{Mk>GR5)X$8vV||BZ=o(D zWZ-Usy1LXMhkPyhlkM*WkXs;M&$YF_L%M!m0KKJRNd4P<$ScS()t7VQ1F1<3lR-@y zA|jzWPVV=);T4<#wfg&AIKe%*q}MRHcE{lj`IT1nV1CYg_GiKsE}DEVf`;(OX?9E# z1KgzIZ5Q>>jmPcj`m^oj69G)9NMcGmV$jbU?CcQ42LYt&T^$7=H+2oTP8R*`O$YKk zW3?Tdk>eyv$%vfT%I?qmwACv-{0vw?W$Y;*dcp;R^~QJu)wamRAsP|aga3G#|8BM3 ze$ZTcw!mw=ZlK~+JhJ?`99uLBGrDQ^CC};dX;L=G zxR+)B)a(Y9Pb_D_a1j$Mf}J%%Z-1%uQ8)d&>W$ZyJs@0!d+C=hd?Stb;ldX0DnczE z57sSJQ`SPB&bAu|dSbu&W0hRRb4KQ(2XL@b{?;N4ta(x5eNAFSA{mvH|4gX-$iU*c z2BSK#zYD0?@NB;Q++4Ve@d*eEmzo#Vlx6B$ZmtwIjp14@6pN64jqm_ofRFVq5kA(g_=) zH2j5ZK*4@5k9UCRz($9!1jApR!?1!alz#D*OBS1TLZFviU@OH}GMN^H1PE;;FsF9V zUQ9U_R*0%Jjwz6t8Z_BD2S~K8{IhDaYn4f;Iuwjf{LQQQR*?LK3#yYg;Iy(g*Xpci z@NPNXhSdrcp)>iDaQ9_(sNq3J^zC)DmCb=nd7q+xFc=Raf5-}aC!nb0_5t{;UfpmC zm)8wFL-X5C_9UB2`p*8uqHl40m@s5F_SFNaAKd=X?AqQ(I&<<5hy!{Q1f6U6XCx5b zPb6iq#O(QvsOR1U#@^pYH#)~o#k#^CwzE(c`faG-xFiC>3?h!pw6XV+Pqzpwzbq#y zQW^IIamaneP4RC(%YGhl?T>ju-fMgvzpaCc{saZs%6SJg{f9GN0WRC)&I`S5yWx0`2Adk&>^jIt=4%k& z;U7env9vx|Np;@)NB_1xq(4fV zHtjHGe04ATgZJmVrLWBDr6VTR>&(Ba;wzvOVj;B%c!gt#PgGzyw(=|8tmL94lf?*5(^)F3&_55ezQ`ms2$YYksCUazQAazW5f~p z0^C|-$A?xHCUyhwl_i0u09A+jbY1Rth|TNUayftTk-)i514?_J!JtCyOePR9`kzER ztH)MXRgkO?sW;ilJ7^MYE!`Ql{Q$hdf@1v)x)e*lIS|9FGPF^k9aYV*#nSdKYq7eg z2`j^Yyn%s-ZiqKov$iEYIXPK}L@||N^y|8d|$s( zy!wEZGUeJ_0I@9j+(LNOb1vh{{}Lvs9X`3J3O{jfBH{gJd)qa#oG18#4uiyeb&Vyd*$4t^6O&u_L%ReIFHZ^nRred4t~&2W61#!JwQ zAAZMj5UP8qFQTx?vOIeY;N}O#ZPfLJwdDX?Y6awK*Q4(;jvvs}5BJ03;kZ8salALD8-XbAIIBHUhw|A?hR|oaaA>p5cfdmj z)Z@A-$VLus7ojqJ3qji1gsq5GtzULl3-xwF=>|*#qW^=&d-JtaYPvFueGS{qc39&2 z1F`)E)%m^>1>9UwtSVpm@ww3bUM3Ek#jKqGoiBBNL{165AHq5hs@-efA0J_uDoh|- zLd0%Fz36yuatpcv+|e}qxo1CT|B%?{5I3G1)bky>3#f+7QwEJoF?$9p81n3oZlXas zV>jLK0!Nm;IPK+X&-rHXe&O~Ah{r$*-m0xta@4$({cI}dx!rndmdMfJ}S*bngq;~@TW!3*S| z;GN^zX>Gl}1&gpeg}gQ@PQ6A2BH97j#{?6}(%(=CTeQCkJE$sK|Df9egh2V+D7xNr zKpBLj5IC@m4_B(ET-$ZgtA2S589l!wGRtl%mn9-HoScz{1=k{{s8G(;P($BnP)`a$ z%{)tg5Z|+bwaaaZNW`4o5vodRrIcSFMQ;j?122e@TU_pSF=}YnaOdTZ&joSjP;;N{s^@QIR=dS=yQw=0x9TtJ}pr+ zoTISe9slzhD0WEXM~z~s`&lD1GB0W>dE1OXpXW2PoVF3CfUwBH?67Q9Zm@q_y-CkowTm zCli(!WPIkjJ{1pqy#nuiM z#n11egVMta73Be{Og>oYNlR0c6#&e&U1W!2`>6^p-zS89IWsp4o6P(6~JqBPe>fuXJ;6_3hRw~ z2(Xai7ht#$F($`G4n*Fmv|~QTdwo2 zOEb);U0%8@LPc~TD<-I3e5@y6{DhBC>vj)vBpr4y2&Vjp7H}U|Xk6887Z`m`3*R{jPr-pC;>V>6-De z?Iqv#;1Bw6@B?*N(%Q$LfY_nB4@>5FwwMtd=KDjQFZiI4JC6WLtj2Ls4G2)F zeCY!LWiPDy+jrh^^Oi45@5!TgI+JVqrFHL{?4%=G?D1`lyLE`>@qS9(#~c3W%T@(U zT*EsU##$f>d6!;_oeafC!}B#h^utUZiLEAJ(F)bI{RY=O1pJ9;A5dtQKhrnpjDA3L zAa}44fg5DFf~$Z9}J zA3_lmOdhndzOpqc%h_mL1Ty-#wZlP~Hp@UmvXvVC(7uR<7c?*0KA_Iu{091ywpO4@ zdd7EADbur8{#Z)72n53>q^-?zLZsp!qg7uL^5x%U+YNgOk9*`AYV1tFcJNk<4CbBF zI`Y4f+8T2_@R)@sZL}R-WBIw(gT^uj<4;Xpt_>Ms~0#T^L`6& zDjSpqxiEwgdkHiq+G7e5edK$&lBlfopX?7S1nE#ljc3OO=LXx&I6LLN)9gLfD^QDW zV%)?y?XB>C6m@B+$NlJ!+7S|3p+gf;dbO1xcPhBsw6Vq~lBc8-{l$ljBeXY4Vu^`5C6#cx5ZMdfxyKs=G9@ajs?$6@L3-(;9#JV0zQFa za>-x!qV7cmUV_b48uHbpOM>)a@*8e69-Muf?;!s?ZkpFM2sWCGyGcfN3UvROU&Z~*S)1&T9(7|^rF5&-ZHj;vMn)vSFeQ=WGs$x*1Sgn9!2 zB2^FAEyyVPc7RufdR&F-E^lDHiRuJIKXpI0na4c@PM@; z9v*ACrv@ocAtlO4=|w0Kycp;b1P?)Q0=Fy{1?#<xbIRR|JKsG-Mq4WqY43rx}`YJ(uit?r6^KQ8`|1DStD2yf?-U(KSf2 zixH|u3)I|Rf8;kR2l%kWO9tTe9$!5PNk3TKK|rkB{jFnJV)6Pi?#1u%=;0Hn(J*}5 zCHYaLF50rg2liX5rJsA1I%0!^cL}}bT%PbjHE`EG$LrBKA{DLU#oG(@+*yQA4WaIC5ZWCJXvGi7F3cR@d7nx zIGlqF(GlB?uL(i{)gEaP0t*UH&ZPb{S|)&GD?}|%%!j}R4$A<516WCZW4OPfS1VS7 z#UwsE5rg9BDv5{XyN48|2^SAKLqWSC3ZoZ8Pmf=a9AT0Q)i!D+J{&YBTnq%JAGYOp zBK5Uu2WgGZh?c7*Bl4PoxxGDJ(I;(I{J%~2|;lZeIOLt)45;7O=P zqdufAq}j^KCE`OR9@IF7I~K-(eGUdbw;}$)i}>(aZlGgNM{)+q;2WZ31SIO$-=fbQ_>(ezU&eID24mLtj!5GMXyGr_L)G zKj^@|y#4aWbwfENB{QHJjOXk!w8)z2s5X1ipcKB}!{@h^ke3r74q&M49&lu^N3@oi zirJMtx6giejl8Twy2KADskqL*%Kr9^siG8oKKR7RgbfHSdbU_AbGa)s6eJw&*{69w zJ4Zig2J#Zlv4FQY^OY^9rGOGdX9n1z#r;_G+2_Q3weh==oS!A^h&JIK$#>O?+K7X2 z?2B8t&fbaV=%j8A!@PN!AVoI$zI1*_$Is8BwRQPBc`^&Uj`%&B-Z8b#w>h`yBZJZ# z-iZPOnkqbu<_H;0)^xC$QZLvU4I{UNd>NmAM30qA&Ea(sPL{Us@4&QyoxMa&K+1#j z19fFZ%YOm=7QX7;T~FIF!5G+d<2za`Z%=~r?F2gD$c0?W4?}NDb(e~VAk)pf&lJHk z>%SL_ngfHLfAx5uX&w%2nMUdF;gngXzki$6RGj7>+Zgm zSI`?~VmE|`-{Ix9o`78|P*)Is^y+=p^Tl1Fi>yB5(zdqHH(yd1?;8_qzaGuQFz#I7 zpBUJ$0mstbwA4X^)Dg_DAv5+EzPnnOr0}EdRgSxQC;NSxBn?!hWN|tfTDaoh+7Fn8 z$JmG+P(9*sA}_X{*E?qY^uo|_1009&F^vB`Ai~ieFUvd ztBi|y<^bt93yaTtQ9>Vy;dKAg$ly$eF0RA;49+E3hZLWr&Wk-~EU%)>MaGS2|MGyC z(6JA^b}}}*Yktf8B(LrDULP4vsDXd#2vSxenqT|;mmD5w?5#!9>!NBqQSb?9gE1#P z5h%}zm7hA?t`nd1sB!VOhw^`*^r&mPj?GOR-#y+<6j$uau8cxE*Y#yeu9#S%EJz#o zh)RR=hlG-4ei!wlX3Xk9u61ss2^-oEZ5si2MZT|-cCSa74}0{3Py2fFyRR4|7ht5` z)_UnZ<4fDvk)ux0UqGRp+nWaUYkOrQ?~POQ5#6WIG@x0-sQ@c0XB->O>v2|%+c02b z#@w6T(K}xACg|<5!-`Iq)0g;I02jzcJQiblRDpH&#jOfHz+~+!JxfDL`{Hk~bk*~1 zI37>?4895}_dMS!I^)$zW6=5OTnGP8o3LnN=@h?57_uT17&t=mPvnD40c-rOuZRJs2!G@wH!>`V0Jgq!J-QR}#0=2B0h z2lhNWo8x#u7nAoNOda4i97~h; z>z5P~-e+q+Bt1z5_NCP{0#fz8b_&)f zdffV}*-C#IotPp5ar9K*=M=YA2T@U3QOiKmQ`UGB47KF14sQyZ;A`-rWN@40daK#n z#koQ|s+FbMvaa*|c?V2}J_{#X-rdplEYi2EHJv{82Mu66oqr6}#SnZ+6uxjM2^OTV zdB-p$Up=CRiaAM(Y1hFq;Wx7vvg)y@&oA^TPpJ7mqMx)}LE)i}xR2&&iERoR5Wn|d z67vNYD$llkGKYeOYXaL}BLB7g0g5zlMdQQoL>088pMO1$84srr`f4g#u!J6TeZJMp>5Lij-G+Hx zDDyt<(bT(b%K}q^w5*y9wY{l2S;Uif z)e7lBJiU(aqM#OWMfqDmdwzSJTAJ9WF%h1ux-j8_$Kbeqii=&oUq8O?_{X0MrxI)x z1)$ovjt)zq=3%LQW3|UwqWnW2>FKR7e9L~emh>B)=mCA_UGneq&p+kiG@Ez@lv*{+ z86x_DK1c{{B_`vis(Ip;Qe1Rv9fc5X{eMr3MAgMa7cP+y)AMd&Drlzab?Ux^z7du z1pL96oGe-|A2XAV`~H}F^KP|x7c?DWS?i(kHs|mv(-rMeDdh=)c+N^rNp)@Yv9<=+ zpG$T}E~=?Nbr2w8<>+YDbe4>pXh8qVk@74%Hs{>&dwF4R<))99m>&2 zXnBRtLh&!wO=`O`hZ-!|XI>kP>Wjg@6cL6zNk!Ib_*s;K@?_SW)@O>v3LO7D2?-}yBqLA;{HDg}4 zaaMU`HTthHe%$>X&^GzIf4TMal>7bMh5w6KD_D|5C68qpUr0uK5nl)5H^l?vvdfJ0 zP5ben#21W3pf(q>DRS+uby7_zAEISNx>Qe_=tTt{?TkCP%uYdJGuZ1iF zG@?$^$>eclgelMev0o0ZZMNp2kp;3MB#zIP+@}BvkBqc(%7G4SRBXwAYo9}A6KlA*-q{gB#|GlBD z8p=yk*nGpX1BwprZ9tK+QNM+{%=BS2p8l0~SZ+X{uYU-<=(W$o%{HzKi9<%Jz!3q( zG&_j5N3>?Ly&7jLZH{2}O>8E7R|xLE2lezIvrh&s6FeimU>9Q|7(Z>_vx}G`;kE!a zsB zZzCB|uD+F6ameJ*bPuqY&_|I$&;t25k-zBk!Yc>Xp~EshWXn*8-aKyjmK;^~GH~xY z_J07tNF+{Ob5 zwXHYV%V;$3PfzU)08BdZj)bAD@y}0i;19T#yl{%QjB6mlU}1M?3ts%3&hPh+uB5J3 zhRz1B5;AGpP9SVr@Gb07AJ-|+@=5R7Ud$f08`ybR@>;l0QWr+WXT%0o<$cqicJXpL zf^UvFi-G@MaikrkdY4oH)9J{!`-g|ni(!5m$(42!qp^dYVT^IiplWrp(Xa{GBZw$hqdih zwvO;nTaFaC;01o6H)~Fp`FXDT>V2o*fK?xCryp?a>p_bm>*mWPgibD3SZs79jbVc9 zZwX6is$k&ejiK_LG`_*Q(gdq?+g{A%Yf6338f@NFs7UvHAB8ixC0-@&RBG@OxRt!3E@ zCtszd`eqc?c9v1kR8@7Y@(tWKZ`;9-o6jw+Nd4!vDi>J6(%0oC;!>DQq4T@?DMpIoOcaH;Jd`N1fC)SN9iodGAO?2 zHoTMlonpIxUVqf1ad;}wZ7Okpn>YrWYZ5Q3N}kS#n^_i|EE%xW7{kPtha)+@EUBGM z-4pG?iV24Gg5oukV?sP+3aqtG0f6QN>7xK`SEQSibe7JtEMq!g16rcpNL_Zh5Ox7M5S2AhJ-+@MKXfEsOMZvm9|2v~y=($mV?VSD{h9>f4yL zj|p!1J%(3cpFFO&NZ5xQw!f3qP+H_NTyC#Yl4~J3oe~OqWz~H-N8%Bk7iehmh&p)* zCP^$)@tYzA0_yJpM^O%y;>*7N7y>xMte~o`LTS0o{1~8G{r$L`?wbVciN7*O{I~|4 zebzj78D?9Xw6yVUY%k^4yIlWG(SVNG({jHqpM1Y1b+4UMIs+~H%JuP%!e`+wDuLYg zm_qSqNz{Y54*b^>=O27=E8pOf|LRl)TJjZ0)^vl9ADU#lQB9=)D8=FX%PYhBEP8QO zZ1fQ}HU;RhlWqRSv)i=9jvaci<_S>M~6dR30HZ&T3HOAWE!Y`GX- zKTYJ$;f+{vE}e%a`Xt18pRP8Q%xlZkHOFXK5dz{_n!oTIP}9CoyGdvp8u7M5C6cPUZn0SC5vgdm+Qpz0!Dt6REOtA z!N=LAC$W48yZ8J6K4I?geA&Y|2N)=v_K+IB2z-c9Cp|yZxp;eH_-8ng)0ls#*P|ue~jHt;4x8Ue^$O^&5Wb4@7LCdxiis5%!DA$D?~lGi_w* zDVRV#!ZaqM3xR&u^KiX=&Tui{VP$8z%@%L`Om&>B^fMIxcuS02I_UNU#g`hWj64$) z=rbLx{ieACJ;ka_L_tG$q-nu5(o8f8#q1nxpfc~(i94LkEY#|M4Me=k>k~Vq&$-gf zbvQNmaORdP+PFG)5%}_QjZd$rjCVTN+Tw0I{t)m}o1aFL@ zqpK{yXVMZ}kB5}JXs#>@Ha)mhL(V6jEsOFheVZF)W(s}Ihq!&+gp$s`Qp)9-a{`Wo zKCaKE83YHXUc)=JeRuBNAFsc_1z^9{{~K*8=$$2sD&1e^Hn-hKQ4C7cM;jCR9Y*zd!|nl%+S zlhwoa0?Ow;a-Z<``8AE7{ClfPLYl_=Wm?A8P`LxW2|S3v*?v;GKBgj0n*Mg*7>E#? z7H3x$>S4gkSBPX!k)+Zu^O&q+132{)&kXiKzSw@ZR9Bc6R%QaG!)x?$KgHo!NEQc8 z>2b-*cF0+PWFW&xLm}6tm$0$C$PL!*dE=AIofAHgcd|dfkBg7rpR?Ct^!sq=Jgsn| zqr#+?ebe&GwE}@W@ue$%=VID!;1e%Czw~Vit`MEec`0FZ<_D=pT<&r zshtHfd+@9oeiUzZ(zMr^FHWExln^mj)p@tc0o({Bv3C9>{NqBMX;(r`Qw3EK&(!J? zlmVF{TxnY|Ic%EYPKkZrcf;@Uxu54Fnx8?>*`d$1n91Y|N^2+%47Ksd5CG$ET#(2( z;^oRuU`9NGz_BLdEZdy|mxl&gzh=73YYnTxFl-Lb3@oM~sAeRuz%qlkurG6XlNQ>; z)<-#N^JQnu?VsLD?jcey$=7(5NScQf!~Q5~r+lkO&zdW=1@uT};8Lk(7Clb!sFF6g z`C4L?B79Cy&6nSWp5Iy> z_%wn$=2OPCGv@cKRilX(lk;b$Y{_u?{fl#MZD&=N=2mPfrPL@y$8ZsnPU#&0yErZ zXR{+U0yS3t9Nuvar4GlBpp@k2tB|k0dKVPV?OOgN=$gi=gI)0)C;%CEuln#TYK)< zA#Jb639R~Or@XlP*L!>riX?jOhJ2bwtf034%>-Ff_Xr1zwh7WAuhl5RzZy;MKqpN9QIA#A7h3ehX-YRVLE6SjDQ!+GeXnw3p1V> z%etV<{#c5W@5AajccKd+$JkRSERsz!{(NGvIZA%1)mwe3SY@=jlvvU)ZqF~;Mv!ze zDz@%3c=c!%j$p&UMG(~aiQ)Poyc}{mKgOSxim{Ua4tU3wk@PHeu`BmtLZ@l7I5X|L zcV|V9jn#^Mko3fYbBQoG^$s_f%pZ1n$#r6nfeL=MgyVMmxRD%20 z)K2VR0T!{(8*!vY<=RlC0xLopDNN^=U3_~H{dlO>cK%VTV4+{FpUD`;H>u9iw(=_!B`a7 z;DaE)B>)d+f_M04JXzvhA6Doy{6rPF15@|=B5hf&eV3%hDQId#f?DaWM~o-&Jl)4oNHaeG0bpQM z`CGix7bLN4-rNv1pn2ERH6x7H+i-#w5BrX85Fp=?)W|?c60%vm36x`8a}x zSH0#x;)sOLQU+TvRvO^%kES`j@_`FH_Fj6dpp6v=^Z zk`r^wB7tp0_pCzSA%AfUy}|IL-k)q`y|qvm-rl}1y5XeMG`jo!)hE!h zOIvZTduSK-qfsS{Az0QkMkD(!M);@8eRw#)+WUR)h0~V4*yBUA;sXAS_m4rOMxnaN z6Bx!w*zy)0a0eC9)4TgV7^v`Xv8C)c^$PVGWc!Ovyy*ve1z2@Nxke)xXGT2qf%P3D ziNVXBq2o`Xb=GsPtlu+27dqJmzK}S88^r`{^`sFf^@PLvaOnF|f;8^O!bsumL+{AR zsC=6Ih4O!30^>zstH%(*X+%P?J09G+=bA@bmTbC@^al;PT>h z$O8++?RV_^xH&TOnx#MiKn~fe^CN4(;zYDXaP4vy&O0>E9@XF7g?eOwA$|6UKwdy&xk?EJd-p9fU}_rWQN|51L(L(A*w(LY-1gW{GWaMHaXgbr~m zK}j7S?QPX7+0z;aX)+cV^cJ(D9Zu(0c)h60y+M*(QT6o!m{>YU^Ld}*WdAl^*Ce*S z8=_?18nX`Pz`Q*t$SP7x6&MwR2A9cj(#ztH*2CC<%H2eMS-;5RvuJoT^0qj>s;V!R zyACVes~2S*titfqNAGTl3S1_q3(0S z^>4S1j|UKtLtH*frTT&S&))@&-+Y4%aD+^N z-qd64oEbB1c3I+tPa^a_HmXJRD+H~m%enP+4&XT|sLOGgc4a{# zE&4To32Rr_0!lR}#9EOzt2w>SR0E5_h;%pbvY=2R!VY-UXhb9hr=8I$Z$6@sO1piluis6(7$Mf~Dqpt8MP`(H9U8aFV)DI^;0O9|A(uV!?B{gat* z^NGiYbn2^u4459*S9*LwnD7k90*4eviWB8CgN04Yer~|X6)Oblq3atLPqnNOK*pGl z!@eGD0l+YX({~#~#YB=0;#6I*CRXk;D-Pgp^dAw#f**Zb}`S>#HvGOk< zEY^ggjC==AX{dPcopomMYAr?WFfQ49j*LV&S=AWZ^CRBMeu?@bbRnIwZ!@4Ck)R3e z2fBS4H!wvtj&Lr(?>zP`?o;;ku#^uzE?i)7W-S@Xq+4biS`GY#J-D8oOLsLT?&Qa7 z&UB3MpEd!&g(TPF5182g2+JDK>iIm-oFdQ9(5g<P7MXlhyrz?6z17nBQ zeLWI3b=V-7e)XY}7gnMdx@JF4(z`sO`6T$To!=9!5FDm+p@^->$gSe?YhF9a7nZ|V zwyMUk?dy{1T#vo*n2)J^Rv+oR%oWD{C_PYY@lHvi(MT5~O53x*`}zq{lS%V33H8>Sa#* zK}W@)MnI68SeQt<^jnc$C!G$LIo^twn#QkWk4oq!53XBW&sKfB`j{A{*SZSSXM5=e zthJ69+Gl}{blTNn4d&V7;o$EzJdiN00Bv#gZfhVFwMquYvokcMOv%l!p&wq2rtle$ zWR$4Th4Tv^*=R|M1K0c(2%7qXb+FTCTE7j?@9Tn}3c;MHh1$XJZ9tl__ua;!K3G?o zedDjsddl_{;*r9IhiI4C?QIHF+BLbVICMV`ct0*3pOPyr{8u#M{CQ{O^3z@H}CrGui+ zcj~|T*$UHi%x_;t9laz?ttNsp^+8zmi)*Dch|KImY?GrI&3pDRHb^Cd5W;nBXnfcs zAwK!*m(MGf;_fe9tiV<}D(4}*6O6Hbh+t$m!n85WM5BOo5^b`By&^a?;kT(rlA+Ko z&4&3D4`2-2k_+4-`z}V!+Izw&xve(Z_)SJFy}mvNLR!JrD(#pHQ~>{PZAjS3lKc*) z30$D$uI2N6M5(u^Eu%9bA39vmp;*L@PIqm(r9TFs`j)s@FEBmc^5G~@{_rqa4bbw= zDpbF>@|j0SYhO~1SSZ~5^=lKFIz~og_n9)kO@gTWE9hUOMS;tJ%N8Gj^4n(DAPujC zVHiPCjGDH81hs76>lmM{cOHGKtMDSAdlAS!47K&h(5PUO=*dc0I&j=7m2VNO``PAr zTjC1wPdI;K>W=pR+9Ts!6_ob(9Akc?TWwdjyfg_dnIPk&I1AF+5Uc~RV zv6tI7+IUoS#NL9cLJRJ3dRNXTh@JPj;)yTZ4!8A9{dD$L)ugy@fLD&jHpILTX?LoT zp#jtgn@fnB3x>8(ld5t))Ir`%z#iZ0Z{x3~Wgi;teHLS#wdL!+MV?Bmw*$rR(c}mx zfoEn;UoMh3?PS)dF&EG%jskUW+jg|DTr zt>My&{OQq*yf)}epdMDg+zWdn1PWz5yC?9ozB;D+Hzl zvOeJPtlex&TD4a7(?Oz@@F_>#*x57p+5G z1WsC1T-X$?OG@v>GA^6Mrk(|KGiH)SzmdcyNncTxuTfcRp5NOOEH36Dl()-Z4Ac;O z5W!9ZT=v@6IF;4#%qDkxs?5_2%_XZ2+w-u0#*IH=Tx0dn;*#?+B;VYg83t?zS2+hY zS&%2MEh0YU72DY3&tdKH(VUK|`Fh;0T09`Cx)yd&U)}p2UJHyzjN3}_u&uU1ZSm{# zF40Dd3NQ96^-0+G8Y^GP%ewQ4EGA+QJW4_6dQRf_AmXcSK;tbKcPW3HP9c>m{lprL zte>NN9wKGm-Blc^J1-(^fbcDxUk!Eo30YFsdg7TLrycHf8Lz|_qRk1SPPy}m1{o^u zTdQng$@UF3c|&D(5H8Q;y1!;|I6}nndOG`VpkuLdAGqRKbfziwIRNu}lL)s4fxzw` zd%Ka?Sfs{FH9l!A+W3~(j~7^XNKa~;fL?9;R9wEoeybv{`%%9f!r)aL^7BRC+`Kr9 zSA8P*>4AA>-|o4=(ee1wFNO!IE<6y=k7>MK?%zmpM;pW7yVU+ec*)vy-c4IrBZOf&pEuG z9^3^5iD;U>WBAT(O0bOH&EZL@>wD$P$@MJZ3iY($TxVDe{Xsf)h|Tr?G6eo+jiha^ z;AU)%(oljXt*CL!#yFx+aaT4{W2yaiBdY1aY)o)R zdfVK;85v=-Q(l-GH#2n5D(bx$VE0MP^1~PH-9I{(v$W|%z0T%Tj8EmxU#+j`DlyVS zgT#O&6?ca10vo~0}>nqBwhQ}b_QNgC9_f%f$dxYM6 z@=yFcL1xVfKBN{46hP()vU#585u$M2`_hoF5wd#S!FeKu`+-$v14*RWnt&MRpKB<% zuv4E9Ohq16r@|>#z$3SXcTadQgFHWMl%(aP!E`!rW?DyZxVyC^!8T6h@7IycpXVnu zLxKYtR6;Oa=8^ab%zc(#dElmSkhmiIZad`2OeZ0lG`#PRj`DLRB@3QIb@c?q;a=E>P8E1-(d+vUz1zRY%+HLsd z{3s0{Zoe!viXr6zf~2qf3Itw!xf1lF;(9~HhsuQk+7@lbbNK&+MNbr z7ycE_dhW$7qKmLQZmUEMdNJS$PjQ0q)BAOF(MA0kBn!<*T;>f)S}Ka*#26T=f8T4% zJ*u<3?w`?zW?%SlSVYNrRlS1bJ-;AiBKlAC-G-Os8aK2pxB_Qf$S6c))s6e^&e^V) zMDpK#i(qSapzre2H+B2~BCP(P)kd`F)bvw&Amg54h1Wj+tw?Uset41WL*XMg*O$~e zjpV$ORp0TdZ2ao_K4LLsNRfW0>e^GW~Vl>pJgqasO40Yge`t8iER57^BG^8v36ODG}}o9hv%e znf4d`9EsbTcm_SgtNW0oKvd6|jI7>tHa{*wV5fQNRLXu4&~|Y>{D~~#D1AR(?VJNZ zi*HhpL^d+cQ>*jJBPprPx2`2K8VwjDI$*!WW<_^{v{JOd=EVns84pBGl&9D?K@&e{ zh7?ev^lN)U0;cOB+cN>2w~S6Gr|ENQmcN6>^Ft3Lvsm)sH3?L7uOEcE)rEmPn-Ekb zuEe$QPFL$r>YFDVad4Z_n#4E}ky?AzM{lH0#0R?^|Eb@Nta2V_2{7^ZtxO^VVmXx8 z6IPX`l2TK90qLc1V{qa6L-qnu-B^{@jJAg29g20$A z75BJH>J!&VK5n9P-t@n_-T2+yy^M$QgqK%k@?h?C=I zccP2iX3X=IEIA|o+{jvl05qhNeYc)I^V5ew*p*vgr)_gKkSx1-h&}MIH>8lsGH?=> zd2!c#cpJpL`GL%`%ZR6`d)(cYWU&q*m#BBsK_tFCgsc(g9C&>6>T7I}%%UH5p0s%{ zRWGjS{5>Dj`Y@hv&<4#PA+HaR{T6L(es3352@eF--)7rjSjdV?k_R_ z(wSP|q$K($HiPuU6RzfUE+Alb*k9w&2bPJCYK9IxAuOvDo?9ZSIZzElloBx9y_*l>Egr1{|*_eRS5x^`iX;|rn42lp6gZpitg*W5|u z#ijixyiU%EH@~T7YC|rXL6Qq7uii%gn{F$55Kot$92hM*hVQ@9vA?>+UJ;NlpKV;q z^Gx+S-iD>8_4VMbpjO%Un<=R+xOQ8gM2+TGS(oDws~uL$0<= zpUgvxej42-91w}=0H2~iJpcr!kgr|<(22SU1|Q36Z#`tv1$tZa5-NrSlgUfzvpf17 zGu_@bXV%Qu(mZTlY^_5p7KTZJX!&IBPh&!L_TCP@CKi>Tl1v*@Z zpCdxoki7>Oc=-3{KAD=UQb6SiNbYuJG*d~Dd?K_A$l*bwz~ zAamm*^{9>dGx~d%Q^3r1h7?Q%&0?AJXXq#U5lay>w?LtX^ z5*|-0Ky@?!lg1XOy$~;3isZf1&v?O}cnvKqQXt}fLi%>7*!Fd>{IY_-7rCMM8OTe5 zr1A{>w0j{cp2TPr&a2-bp@BK|x?^k6m;JP)K_`@b#}EFxpP$isXkXf8d_<@HpBXl8 zA0Y%$UB|x|>xYNoFE1DX{a{ybuc~Pd{k!_!Ip=^HwYvkp5Wlnj(cL2GrQV_%x(cVniBcdQI>r-= zSN?p{2<~h^r{r@QPQhzhO#16cn?=?m6DGH&xVIw$qhr>&TEsKhz7lxo5S71u+*iT1 z#D-rB#)9i$^e*&L1ZTtB_f#9Sk#dG4nxZ_f3u>HuB(vpXo6ZPMy12%I5P38GT z4&S?~kd<~p+I*eY5WuXYsy@M*i0Zbk3li1IKXf&!9WT>lz2n#41oMa_Lp5ETvwvKM z75N>Oofj0;!s4Tn8r0xyYO6DZ+oJ2lS-G zDBc5joPPMo*{9|;iskdc-2A@EtMPIjl@mZ8pp69JqUby)0Ot!Xk6W% zxsK^|;4rfj(wTu$T6Y&SaS6P*d5P}fWig>|BZ@eG)JNF3C|=r&wd@VQK>?qMK-4_m z^#<< z`lX!5@#SA|B*kh3|CL*0(hP%4*Rw$6`Xhb(4TJNGqgx~oV0ILb{JmZ|kEMbY(4>+~ zn0qtO;yrzb_;}ao+z-x|CwqkF2KOAwtUGR>;92ylwNcoGtnzcK#!hkB_~xb5{WzKc zH$KhwTTlO{Ru-w=N-@|c@bMf&NqvPOkfn%s2&2=>JG*M@AvOI8S3zFGt)szVDwZPxJ?HW!=_UN(#DQ!>pSu}mCci}^vqZZjvK ze%c;k6Mto_2tSh8WPp7mI~&jMCD(w&pRg=e<=j~6h1fxrB88QPRS?*0pf-YkM{8Zz z8kXwMrx87;gx@n2#iko4307g7=wQZ#b=-bdNCe!j#ZZg^Ybz-h^}RQ>(AJReJW;WT za#9Ar_hQ#5jYIPjb{dj9a=_1|Dp^j!yUMO^c1m~17c!2PYe~N}{B`|Hl3F?N@kE=! za%SDH#y)(`1}Y`(`7KYGwBX+Yh46-pO2P8U)b_o%9uVqJ5YPjEo$e<>$ z5Lb-k#hM{>!P)?4RC&VZ9uY`SNYSv=WWE!LkdA5id_UiDeM(}A#?7e0uPa`&twMm3 zL$>##qiCuSG=}<#?uZ(NXkX~l{gUbL0&|7edO&6z=-##){{sfVxBf@PM-fHUBOdfal z;0n)`__*9V`jv*2-WD4Ql$TyrO)$j(UtgZW#vt9d2W(cfPE2xjd;$ER2q|`;+%e!8 zskTq@BRL|3FP?2>T)6_(W~6tn!-19`Uc{;Qee+j2oPI5GFEv$Fbx~liH19*nafq#9(-HOI4BIP|Fp|QQcUu^v}Z)uE@+A!v){LEXYILF#>TtRnuGrX6YlEs$In0VCHux)AYk(ia&Sb>qQP!!i8%fGy>y9*+>+dJ5HRt(jmwbt9)J|!Z`(ufJ){PN zZYj`j*|z?*L-#dZP-$bo2K9da-#>13UG}aQ_E36N9dc8{tD%{+-2 zlBPs}bDbYvZ;SYlqV{5T8M-J^11ot}@17_A2X@NGBd-_R&7{L`Je4ARTS#ihb`F_s z`)$5jmRaxJ@oiBz{Gm0ZzF&|0xLz`yz0`9B@h;PdwBK`G0rh)QMUW_BS0|#QS~~O} zBddaMZq*1~dv!6~)hLOI9GqA;i4KIQ1jbDy9C2V?^iOPA8nUdmWo1DCw_X6KO>7H9 zOFYsKQy@D+3?-PeBM213b9Y5`rr#n^vZ!>{F4#}|>T_#cag_ejHrS{lHdAiySz**} z*P*v9Yk9;U7*vA3s3iXQy%6sUA;Myu_P(^Uy|Su#Jtqcj+AM(FD8Cx30_&9*ghS_PXy3{PViRY_$V_2D+8Bb4oJITesI?z zK#_iu3H5w~k*tsOinpTCIzEwym^RK@FsSGs@+l9GhmBQ!#ufswfB4RpxIWkWJ8VSn_Gbpz%PyOsk*}=o%$3Ul6VtSf#B@w~t&nZo)y zOPnn>gm$B$AKTXVJ{gMFVqZ`2>*mN%uRoMly65!q@n=!cp=0n7j&nNHpo<|YVZZ~g zsgi8#x%ut|juT!W6Gqb=E!S9!MCcF-4%_PldedQ|33W_O;UL}appLo%U81vZf#HMG z7j+7dfEYl_2lXvGeY5=xy0Vuj2Jds9n2~A)W2aACu+OrVR`l=c>h8-U7E-RfQZwTk zzkB#1-{(&ztd;68Dh?rEjA#^-@nrn_%Mj@z#8fG)esp`OYG*FkgNkCxbr%%Qj*p|# z$A5OU2q_n9?7&7CB2eHMkmb(k0>FEqmmv+%F1nXraO8Go7Nn@@Z|?FP*$iX{x-d(5 zn}UaeFX0#SXV@~r!!1-2d1jLj z;D0&~TM`j?`5sjUKH5=L=uNdh>f7o6k#t^NifU01{!4&hcqC_(489>MS&~Tj^;7h$ zHS6A)o*vr(=j^k0)mJfrrO)uV{4avnOLfR3?0|&$uO-`!A--I|$3#Cmh?T?yYyJaH z?A28E?a|hTD|}HkKF&cI#uwv#ffP=Qic31^#z>>?%@=J{eFlptq7I ztWMCn*W}WybM@=X{dHc~(WHOJ`{hkrn(n=i9qZ4I?AfM^G>6MmKEqIeb8u}SSonys zzCT6Z{pBF50YBdC(9T1*naZ8Kr!FqAArVQvwkV8k*0x;e6dxl_7ny`jH`p2+`Mf7 zwC~$RnJ*z;4~gH9WGMI3=}cj+IS)vWeWb$Ub`$otVtaP!;cQ(#3(^k-E$hn5ML*Y4 zMcz+XkRW6c&c^#3tdU5%=hG4f`WKWo@z}K*JZVT-{L9j)^m;jeY}K62Zv2!y+EVR(UAfaXck90P;1}0Vx=m{*Eg|5gg>nCM4EKAqWDffdDb;(La#?UlOceM;+!;KW-|6OW^?9A!j+=yk`VWagiIg70$>W3S!4@)#pGPxpGT zQj@*jx{a@7-^`F8{RAit9&q-7Pn~<4U(MMmx%n%;T2X7g|0A#Lj|OUV+Us7j)e}*$z{FGpcm)+8Y@iu-O|4!duhrpcvw>-W9ch@7stn^rE;ra2n&{`zU z6$Czzk!=@&#Bf9(*?g=hm03UR<3y2Y^>|TGv!GA ze)XQ}e7f_vQ1vNl5eg0i8m3v_diz+mGt|lQg;B>r+PJ+@LWjsy*3<%zih#e}TXR zp7di0QOqeUQdxeal9!9hfa%5wHVh7 zW30%1zr17QP4vpS{7lmJ5j0R1PP)S+=81vsA*&uZEGhut4o7j4M<(vp?!FE>bKS4Y z{I_*ME&(PQLk{4&U45sb%rg-s*m#HPlgiUTU*3Pg$;SOYyVVq(Ztu$t%uyH(jPLb# zhCq{aL3#oS5>NB3n6@n*pmBbhukpQy&w`yKP7>3V*<5<(G4f;1C0v=^_B-a6*0b_Q zx(PtHqmkX1ggbm5k8)s^BRTo=hp{KunS;CI0WNN zQkj8%GY|7Tx&v%GmIa^w3ycr|iN8djZ?0@2+EVp!8G|#} z#?Et8b(&d1oemc4IKF2u>0x^2 z|D649f&!0%-Yk4C=X7t;iK+3qVs@Z=RV=!e9KziiYgj(OZu&q6YV(m1AwDL=_uf)W zZHkh{v>z#a<~c!BK(Qr)fXy7QF=oIGSZ)oq2cA~&#TRsfMvKKCUf{^=+KC&e8H=fVrP-7TH>H zXL;+Je2@OqfNM(j!l$~hXTyV*<=5FB zts=GTs+PVx{8*^I+AlE-9WD9`fcpru!dc)^)6V0o+t_z|z+@}ZqNhgLxOE7oRkf486Vd27RH_K@!tO?s74iAui%RO__# z(X)(sz}>j)|2_w>3Ew0Yh;r{jusI(E=&_LZdh(ObmmlJD+M0Q~C~RhN{2~x$edgOy z2!GD`v(;JAw1<1O?_Zi>z^7B#wb@}dx)<7g>;{;s?C(UPe~{PN+=@mLu^E%u*X}*! z)ylJan@Z`*dp@kCf#c_zjiGQUVzv&7{`rm38&V4}&3{7lO@9jTmJB1-l6tm|>CC(V zak$=Z5*!yzh|R8d7k{veEAx-2-l(-+p=j9+J7o zYdkwXht663G@z<2WXl}<6LK$u*o6F4so#|bdUkkt+|VYF1iF7oVTpp0{)z?cg6MHN zEMEvIrfL4Crap}bzJ}=dAbIz@!C#nbn%2^v}?$#0d1Ne<<#Z5^*Uyq3IQ>YT- zW+`V^o_Su_8KlhG;?4u7;0&BLF%r9GQS6>e51{x5Y-K*<`g4e}MNOdB>vz+>bpdG@ zn7{dypRb}=IROhk>wRcIv|<9`*nGEOtQ&B?ZixhoZXe2LZtdRKl{t3bn#~Oy+Bzu` zI6)UyN}x)9;tqCfEBdv?Gcep`gQNF)9tRG2VgQ{ApO92i{raS35>tULj%H}{YJ`Xr zDUTUcF{3NSn$OHQn%LpFjuI%}ZbEm^f!l^pM2?y73SLWJ^+F{T^;y@axU`_A64*N7 zD^&$C3p*YSfS_eBc2{cYdW+)>-dO6$xyRi3VJSm%+ND4 zE3r$VnE|gJBGa=A9ustWT%hZD18CH%Uu8cKtMU{YlXHJaUtjJnO=I|b@i6Bi=F0|j zt(7{!#)4UJAfGa(^a1q3sv*662#^!W#-Rhkc3s6(0ex^*0sC*A@1FlRHasGa)x?ra zK2MW4zBW`g;|zYy-`V7VgnDE*HGSlW)p>N>_MzR;`3<0PaCVkZ4>%xOc!6nI-KU#<>ki4 zZi@q+syL2hX=l^K>G+`F7{6+0S;QGd$RlWpiN)SR_@#7YjF%nI__rqn$9 z<>{63+Xd_tTfv6RMGufywkyx3K|9i4#0^W`R_c(KW#mI{z#~Skg2ld${R?>w2?YL5 z18#F#DF{iuado%5r}(Tp484Pg!<{|oqhXkOOnsIO;bwKWpX$Eocy=$dZNT*bL;>fV zukf0G=q8)K{j}M~E4+a+0xG2Cm8Y9}PyNw{{*cNKuW)A5?`qxN)J%*aE^n8d8cE|$ zi{v3vbAIq>eT_5>wB$^OQ$dLh%K^@i8iOG~G!3RQ!#-s2_w}jyq3kZLuJj z%ikNs&nW2p8ZE8+K1jjh#o=s1>D>#ap)^`g21fs3Y4Xr}o8Mx&-L`$XW`$xL?*RV= z2wWxz9P~bg#K+6{XNHXT>jxA#Do2CJ=R+F+DVEOFB@I6cahad?% zt%A5fSV-Rd^sMs<0=bm>ha~Q_lyG9rjDk6gG&;D#VH$!bs&+hd8GpbebBkorSsi+p zw+RXCt-fEL3UyT+ErD3qVJGO?S^MZN$uJ3E1(8B=MH!H3*T8g=8{9a!)1`v^-5^U= zfPKWD`r^3*M#0^~F7SW|Y*bSQeCOftJ3p-cB;)KObxCvcwOdIH7VcA+V_~&X@&mu4 zb=>nE!ERtFM4Dy5tU(-hO&a&0yiB>a_v$w}*Ht;3O zFKt;nNRNYNT|dOK&wz~uG@KyE#CVIvWAG&}BbVh;)4jj%j~oSc)P0$dBEVFa?uCQx zATH?ScTlc$hs<9+{Ru6Rf(!atAovCex#*xb6QsU$E^p=`N9C5RliE?w5Sg*cfUKm_ zdgDYd73a1zr_mukZ2C{=Y?UGU-AQ-Yh6~K{@7k9*=}`ri5Ty9{Dcr^@2E(mYm-3 zkV%vxonEp@*gXP>d~ixH#R1`5#nC@V<>AY%&ZiXi$;-@UThU^Uj~^ZmmnVD_-ZB|8 z@{Zi@k#*W>9e|OAR2el4j@C%NOtYqY9)Wim7a_RkBhpfm%6XlGYuNQ0h>hRx*Yo7} zs$I=ThK7lG_j>!KC|zSh5CR=6aq6pYpD6DS0}NLI_x8J2bH5`v0H&}JI_bSXT9<@K z6$9@k;NYUxiq$-CVX__XeL_%JJMh7-6Twd#>2}`UvYEN<2QKJhTxISEAni&+`(~Z) zM~_qF`huClkVEudnYFpEAYXhKNKT)52i?B$dh+k~>gwu<5u@JHgF73^dl$6;SA?swyRK1|tzIDQ#?S@qxpb35O$!?D(UUB(IVfm#n# zLSc72N|*kND*8FTkb|$5rK*10+?g5TP~q=e{U&1nq{d zWKEU~v1;($Fg58RIz%Ekvhszs4VpgBFPoy4 zg}vi}5!8z0jV_}Z#x zm45vSto)&ESL{-Es0{a_-P;ROvOMg+Z7LVu6=hw_Ygp4s3~Iy)i2maNI2~%kckGw* zEJ(zc-juY5V}=9!_~qCP5W``#<2q8ZwQ`8C250Y^TqfLPOg zqnsj}>g?eRqvauF;-7&wRHnGEN8J#N*1of?OfG@5Tx0<8{Vk9f>=n=F>@Teg}~; zQ=q--qlJzju(XZV0G514j!1psy8SSNU+fP^S<*YU1qME0Xs z$R5y_owjvyC7&H-;0d3^VT6bq@bwQ+C54XmG|*C7_uQg}%D-L;itBzE#O;D7=G-I$ zB^o#P73gn+a}9?H7BviqkMf{Yi2 zhxii(gvFr~x;UUUpr#vwD2Fezf@lC4!w9E%?bWGk~n9mH8_2l%3jpPo->W z6{S$NU;%Qm4OQl|;lbr2aRECs4*N-F7<~e(WevSj-Xo(&caT}vlFD=D>ZC=eC(U2= z&uW?2y{2msCRv|3cKN&N88e+dU(Ohj5PWsN1IgGwKv#cv0l?T}_A*3Idp~dzgnB#` zivn|;R+^34*&wOe$8;cjqqI*MOfmEx_GW76VuQcEMM7k-(0}z7o-3&?9N@tl$^Rzw!sj&gU=GHa&i_Wws^U`N||VNFs}b%Q1H&W5UekX>+R2 zCGjkppj(1qT_J&F)K6?C*i@QjCNR!^ugSKgd1447@OW)aWR4u=O9JMG$aVnTYE%*c z+ln5VS|Q;ONU?la_hXq{BM|t9XG!I_o21^}C!8+<{r4+uvB7#K1Uaj9Ix$yY%vFG8 zo&+IXRs;xIK^4&AoNav27W?TIQt;DORd8Xi^u~^n%X?1>b(GfZ+w<5$l-TXF%!LU` z)d}Du4c3h-meZa`Z%sT zun9eDjj1rvDLriBV#z1<{t>y{^BpKSHT(NUMedyfw}y1=5}XROnomvsp3|uxaj@uD zbsZJAkpTjm3@y#KS{3*_9c}(t67L-9A0HMoAdYftV88t`I9?9QC_ErBVi#6Yb4THMYE!Y_#Ps~o?N*@YUizCTSS>>!@q zYt)!BC@c4hN2l3WYt4Y9jMgCz+xY~L?Hq2kO-CZ5TZR+ctH<6UF$IOM16~^myDK>^ z$TEFrh^mPZ<@=Ex3)JddRsBoEYJz?*lP|C+>7#=y$i8lMt?Lc68fD-aPMJQbJoysY zwPNfBw3IWnWtkmFoSNPz;(>G%G?Ow+DJg$MP7cTTXQ5ZTi^;s37gvDx{R2@cX#G4H z>AUML(hgRa48DW!*to(Hf%o)19XdsMBd-z=;cnvnxuaS1Kx=bmA3(TxKn6E83)dZ= zRv=p=>4hLUXvat1t!acjL=D{}%Iul=!{U#C9Zg6=IFVe7by8 zHTg~Dc|#x1Bp<+dql>F^eUJt~#yO=)zIf^HZyz)mZ zDIDy_%0uMrwJ_bn{PHdiC8vrUQ1k+9B?nV`IlzKZtIiE_bAeBy`71LtecDgUIcZXz za50kz{eAVkT>F5xgVBU3`FzfY(yF%r;brqO+_KcBA>3G{B`%;V!=nQo5olY^@;WU`d!WYhz#qUbM$Rx0mMtpS9HY4 z*T~7Z(We_TCaFhVk7?!lQk8{|kr)Z>an zexYQrB6uFeruDJ@h`I1jd;c5zB)iclzBuvzOc3`KgT5^G5B2-aPiLxDERb3T;wpk`qMY#6GSYxJ<(omg_-M_CXbKw zA`@cc2N-YMF{dAM!z2|}zz~Em0u9-CnL{{N^15q#;Y_;QN|`hlXpY3h@4V{;d9-=^ zEF?1tFQ}R$)5&N?y|s~{Sg^h#O>Twft<_G@-2ia@_|i9G+Oqg#*Rfr3JQ!*9YCVE; zbEI}Ll|e)PRktU{9U<3Ce@=g*Zq^qu)tRfKUTnXD?;X}|cAvNIcyYN8=7WWUE+_)q z;f2>Tnten2sWe*v5?DD`fPXl=4PxF7W%nmTnDHF;njgu0TH5AajD&kpN9GFm{7FT^ zn`!`b-+BzS4Ov;RN%i_~h!t5;LE*QitM|5r4-~fnst#Er0F!&}MFQtxYcjqPJYvw> zVRi4e9JQ3xU!cWXQaJLmM*Tpz07CYuhV5ByP9qrUXBM8*WYlz3aa*>c*tYIj5*xwc z5F;14`5D|r*{2jmF*yp?z{(A6gm|spK+eEIdoZsvHJ@5TzEFx7HE4Xgc30FA7=?>s zbl>+B+=y9E3C%6i1WVOYxSp`9@lipj`}jT98thf(2p2dx~GiV2u3S zxyWu5)J#`f+>KzQ z^!jpN;^P|TCOb%0#~m%g`fZ~n37G|Cm&3`oZq2?xyYxee=1DRavNR-4PZz{9BZa|> zkP|Y_r%d6AvrFSU^K{O7`_zM#dfXRSx%2K9;l8@h&ysb8S6!og0%Js|S!r{YUf$Qo zA?U6BBH}njE1CUy?kA7?dNa>ZhCr`(dLS+q{Ri8HIAPG_!m|ZZJHO>YRm4x=yke%u z(j~dekzxRZOXxSB+ihyy(=0O3Z5`e<2*Eo<4lL_jjkuigGxENM48;%@n}A%rz7UOE z^N;7-8`@$2WYvXvcQqAB-pILPRYl5m|8jYF2*DyyZ3Np1!t5ViO8m%Qd$C2Av)tIj zT>kSy3VI;8diQwMjtwp-BO>hLs$RY@6{{zNVcLDtZozIBkVuYY9ymHp*UdYZ!9ryx zwh881cf=j%S2CXz9ZT@A%7VfecZzr%CZH`!tS=6GbirGY;aw0@F85B4;HPhn8M1p$ zRC752czh$#8j?!rWiI^Z?=Au-0$*Wy>CDe4n>kL#Ob5{ip1U`1%vs^Fl5OncL|!8K zCX-L1f~05te#>K@Z(74Pm;`T3cvHM(b?MT$sVLsRwQrd{V@5lMmELu9zLC^Hh^Ru%@ zUba$1=T-OMD!h>HjDI7`CG6MvL_(PDH`0^$IUzgaL0{7I0x#r!RKYoF%o{65W{;=| z#O3f7p^DS%xbeC6%{NA8g+FA;Y44!nJ@j1edjaldxvW23m!2t zO{dMrDV~%U&3^G^b2pz`Sew7Kdp$?(6`}L%RjNc_h4I%^P`<$;1r9UE0rEl?CV^!g=Z-{dUd(rgD2rFkemFLT76Os240Yt_Tv#?0=#R{ns(9cCwP z;KN2cJc!rW2RKAvH?tCKvyZA9h)l@Qf=|@hif6AM79A_+^smQ6_~Vz+L4>#a&T7fFs8q<1uqX@x)x9W&e{RA)E}Ke3l}88 z$P&_T>#Ybt=!*ps>FTg`>eIeGIW$b5Ca3y~pMwZ91k@Vfq&~r{FoQ;ClhgT%8Wf#x zg&+KZ^DWKAVG!G;+rgaQ#n)$pF`RHOaB6V2U9yxV>I+bEEOn>{`-pxFSsxR8`{l4FI8dYZ_)E=%<7fxEw1h zCa$_1{0=*U^&Lh2ZjnJxwI_8*URR_I7|wnthJYh!KtZ@iydt-)zxIVEZSQq?*Rq%^ zUr-Hn4h|I8-QUUQi+R!U=z?{8EeRTK9OHw?CLdw(En-DuA3Fq)%zIr+m3q+`OpTnL ztm2$s@{J~0)liJc`o3Ie3ZkapVqszCRNG_4w%a`mHK@kojGRtk+Rco)R(;F3BX_rJ zv~#NblXL{w7$I7Sf;u`pz`>;%@ug^`3zcM~)hHna_ywO)&^l@0)=hwr1lI}_K+r&v zxYLC374-(6g5FfKk%WNo4xI?iJ7k_$AwlPh;CHGTT}&wvr)Xm&PP%n(r6->t9|%18 zrt8PyIW%H0)H-C#3*v_?oA|CDX%<4Pn{?;bQV&;Tp^*0+_Nb42aoJ;Sybw-%O}zas zTSL89C_szZV%`j$?~bH)ukYib++4ST#8*OYeVp1x$k$g8m8|PgZiRzQ`G0);i82tB z-WyU#1gWXn%M~`1uaayh!3)cdyULg~g4kcX;hL#Alx>$jjG&a+DF4sjWS< zERtF)mIB!Sm&_M2p0z-Kc1orn(MY`K9?mQabgXx9 zMYXfJ+_1wwhy=ZCVXW$;W`F^Fkvh(J42lZ?mq4O`W!l6GxfvA#x~Z&*kCk2Pb_DB( z#2sry+TfY!%^HK_w+81jQc)2WErBK&Z5o0#5*PE>l3*!F=SePpLW_p{ZnR&B>Za zhpuj8QLz^oStI!g|_`ekUtEOR z9lU0@&B2s%gF+Nzbfj62ESML#m9%MZL8INb>e&3+)mh)aw{)(S~Z zz_6jeV#5+jUC30M(Kq-Y-rUU1F~%YBPIqdWz~lCNmgV7=DKp*xHI-{b+u?VGXx&j{ zCkonoxW6UL3~}g^xsPR+4>paE)XLq^*u0qgOk?vG&J&ciKqgKyg%6@N<@Uy-*X&j} zX>3^@K-V>)yyt%V?x*7s@$0ds=vQOKjU@hoiMa9k%jTM(0QAvWJb?fOOA%O~{$qG7 zjK6Hm_Ip^b+g|8Yk$IFPRkSK0HM!)gC%AT@NeZ!Ff!7<|fBSRpK`Yz*5HJ?ef?!U* zQTqb^x}*rB!qfA!TzP~3$AInhKJH6Z7NlPY)gj^2a1yrNV=|KVTUnLSDl3$dYtjA1 z$jsbeJ=T380v$c*;0r|xy8EOK#~U!Axzz`hOE+M9L4#%rE6$0?ZQ#~5MdBHaNa)Q& zjG&f0H(o^XC}uS-p~KK^3%A%t#1@@s>BK$K0fXHE1*Sd}BFx)tIZO9Dc12)YNc`~d z^T`|xh$i|Ygn_*se6t!*^>T#rRIfG6;lTH$KvmKNiK-1$jJLfYECFxLXO%Usz;MQ% z;SQcM$pNA4Ma3rQnDA*zsUtgCJ3_;05J>se2DB~S^w;!yQjYl|`32{VBm@Neq}1Ei zpL4H>h__cAp-rJhbWV>5MZ6^Sbb}&6dU<&%y(=>{>KgTiTXg>WR}BxO50k{jn?lQm zHXi~h(RRDO10N01KK<)=liUDV+My)Ug#2Pa9sYUOM|fXN7X@LBsP5`WCK>TKhkps8#u-OmPFrS+ZhHoPB>1;r zrfd$b2_3TmrY)pZmlh_cbr|gj?s}+i8DE8XiZ+pzROV3kYWn$fWv&mm%x)nG_ls}2^Py9#u6o!M94eR^2+f{pEB`#m`Q;pYhSby=u`jf^_IKIs z7rC-(WGuO04-+Ji5vNN*Qkn|JhWxXBTl>1ExMv+S6Q6^iIY}2>j2=7A+>s5)cb{se zn{p!68$!-hdeCq|T%M@$IkijsQq>&yyLb=TOw{CO+=b1r*g5W;`79Sxc%r~6@kuw4 zY{nPiDz0ZYETGRlcWbSx!ntsM`vLm)Kw5(A;;7!68w=>_{zU|D@aHoMUdg>4eZsHR zJDRlp>*L2&{%j^?U)%!f7M-8(NP)+Z?i^>)7{#L2Bq^LbVh$)*vElQ!Fba5g$8ZUo z;10L=h*&P?m0vfUF1YW6BwCzx^{zIL&*3-T4De6F2>H+mh6w#ZmocO4z01Yf*DT<` z+pvF+vRVDLdTyzKu0yTQFf${o>NN2)yy8h_h3(cerhD$um97b}mU>xdeXQGL1M@e! zo$>}jEpZ$Ta34|ir(bhZDhp2usYdj)jXc}<9_Tahe{~S^3E%y=-6Ap>-!bFt|I3b- zk2AkTEpCy&``LF@0oC1DxoVwsG43Cm`^YUj5E6QXMUW&u6ft;^yAF@P|53B@H$Fee z;!~fgqFq55sze)ORUH35=0!jF%2RDH{~Vx4MH~$~DuO7=^Z*Vi-g7lo zO1ix@1F?0^a{l;LtRS#iQ5S&lm9S%jB%Gv c4fDZhW#jJ_Llfr&F)v??L-P}3u2 z^qAdq5N!%BnX#?bjus!M*?O=bF$Hgv+4km}c~*i2Au^i~2)TpZ?S-Sl39_4%fT6Vm*yc@NGwbY+n z_I=5&`MJDj2Y!Ez3SVJFk+~XRAXs(hBKi09d*-9Ll6*Et8K8rBb;zEj z_5GZ~P2`Z^f6^V0V};Mb4z$I~$$TDP!x3Pd(Dj*d&Zeg3bufR!W7PAO@i5bFayZI{ z6N?`})w@#jt2spgn|ZCXzZG|1`N>qUDpZ_UBfg?{pbFyutIpaOm+Bb@E`NXW$>bEL zol$9bHU_@(J~Pz$eM?|2=s`!f8Nl*8?tUcMBUG4g8Bwv(e9a@uNhl0I`+)h}Ca!n*2&3DcaF&4xrgT0G0yRDp&BR;L zxOsx6&|N}2rH#f$hxz`AGDz!%9H6xBuihc#3syl4B{4=L`L3n+#XgR!1ZP~*d3934 z(NUd`;D_G=$%odVe`dlxdx(Lgb3cJg!}zAX*F#!Fm>>YwqR&1Lcpv2Z8k{&2Gd1eq zt9jg~(;|G_i5{uC^wCrL=&b{G!Zvx{D_4bl16GgJ<|;HOq&HO6$7k73W$O=Vh|p<1 zI9ta&(+1%Li|fNN%f+AIf-F@8>Y5k13dfw7blj<<@eZ9ff$qknCwayd@)leIWc0Hx8v!shD3_>j2W3<2kkHX4sQ>sFka@ zd*#47UfhG{*=*d+XSl`pnD4Gh7-V%EUYSuWY&<%qit6jWf+!Dg*LNs>;CE;rekudO zc!vdsZosQRwN=y)!T^gHCxmhrbJ#>>qe~<*&?cL$@XTIFKDn%3-p6*apsr-fer%eI zctZv3U~UN+By^jGImgOrzxGO}$EWjig-_!noQVJuzIEffJ*(efVd_ht=5PDQo;2UX zF2K9|>e+R_!%I1tL%edVBi9y5WaDi=ACwo%47%y1tFwsg5zfNNGAo8(yS*cXN7Uu@ zQEgZ6aT#x3wXr$kIeXrQ->QSj__1{x+8+_v;A;js_WWY3#6n%fanjkBye^%-9xLaD z|NQ-JxTGZ_`$kV0P7W^$Yw@<_CPd41`6B_T^Cc`DbA|M{z#Mt0d0s)AYhmBvmtvs$ z`-sPvx7 zL{0jl;LSq*%^AI^E=s&_iy999&HW2)1~MAV1K{GGeh%l*V|JTpOV`^Y7@OcC zsmQKe?RIuWWOkSi+AhdNR9fSKQVGmA8EKJWNYovNm$zxu=FcCuEWe`+b&Y7)jsK5?u4XGZ{Yx=R*i; z20~*!0FbIPz%d-A5owz*D_JwdfWuk21f_>vsh^}O*10iI4hSbWAe6|kOj&Dc|Hid_uNVyP1H^PO1^ zU_KUK@UQS9%ohf{m4hy!_~MU61N2G;jr-G{a5<7s070x(PI8W+FL^PK#80wE9TVDO zU6i(i9;8Pz#O?>S0=m;y;SLmi{yGITzB;xErmSHU) z8=beaB)eYD(xjwF=|PVLlhu2xOn))iDs4Dd$t~)xcxIqLxHi3A@q>nMw8wQIv}Z7u zb|o-CMLQa!VW`9s4nP?*{O01gRegEleoi}aDenMT$U&<{`z8a3f^+yjodJ=<&p&Qz zl!3NfZ6XD3p$YZB#kT=CT+akW!d_;;mR;1S+F%Xd0|w4hJ!M$c@+c#30E1Oaa2lbNMR#SpZ z@Z(=eP|=D#nf8i|Kk&RER(>51AvW{lA*RddrZ+q~;|@P)?rD4~CyhZ^ zIRu2weX&~^y6tAPPP{_~r2vfUW{$ z>U7@GveP3q_Yox&x=U_#V6z9we zvj44}TZMjsXTF+`P3(U?8N|<$_%Zv0x9bZm|HG~?)Z7G3;kgcqpBF1@J5~ri#$Tqv)qSZ zJMSP@GJ0`bljqg9ep)Gsu-JX0WPih67~*4p(juO{ixZ}N39pcQocBh8=7LwS_mQxF zqp;lZi<~gv);_63LZ9779c-E{)=Db-SB5<+gF7UpCk1@Hs7Li4;`aTbfp~n}3b@jt zdcg3Tt4+K103K_u{(s%oDl0FUvzvQ)iBRMfwP-bVlP8z$k$luce^6EwSWu+jwba8@jHVXyjrjDgeYgDr$x zD1%#qUrYL=lo?G{Bgg0IbS^m@aYb`Ck)S^hES|GJY&L{oMfLq8Weu(*Eqe2+3Eq+f zaKhTSoddA%?}Z+~d?SHXbtjHP`QshELp96CxNolqZ7TOAwR_9;OS#IoT6rDs-bYb@ z8;t@MIqnIp$f(-e3{6dQ{j}_5M#5}D_JXtQr@qFZe4L`7h3J5 z9VxVZi6OkYF|w9yBP`BNae}Pycr{wM0!_GwkjLrI2kUv03aoS7YhPaJ5bZGSMy1r}8!-rUGyX9zr`UZzAoDSo__+0({cR8hu(!!ji#X-q}#j zQ!7rOi~MN?pQh{J0gg*Xr;$_eWs7cW7?p4lc%#z{l%? znBrKhar$dEiI~J>y}qoWD(A=JBYq+mU83kQLDUUf0Q+S>I<^A^M8+MGhT`uAmCwi} zmAU4~(ovDSzFJ56yR^KvXsYWkXs6jeD@vCIVmfX4bkscp%)L-6OADYCl@@}%+?>_i zg>2jPwo>EAnLwpV$@yfGS0ol#l8ulMBt@usHQM?WrWo5o4uZRhc<<~kX;uB;pIAfn z_iZib;L;ZodFI@6ns(vuYZgdXfFP;j0v8q$NQf?W`+F{*p4Z1sD=^&8IaK}r1Ej;n z$0dw-i0e^_MLj*crcqL_Wtkr(75OXv4nHHhNhnbD6@lfx%ky);6&IH7p|{m@1ApJ= zUD|OjYcA`+vXU#+$FDuw&u8?OcuQ$woT^bIO3UtRW@QJJa!g*6eBB2~e70yjMqflB zcbvbUgg$1btyv!W{MRa!l=woEvwTml^}v0@OHC#~$w|6mn&8hpGDsZo&t)pIV>Eei_!}0 zpO62WZbV8s&2J&`n3n7Co0QgNaW?k;G@3-~8ez)k3;EKRW`%#N$YzoSPyNt-KKpgm zlc@^8;G9>vL_ITvfgccy>Vd(Q=MKt4DG9jV)}@1D-ySe*#%UxX^||*m6KUfZdXh)+ zydq*#{=oBX;+#n0sgC(Q5H-ZCcAF3M2oPcl_ht&(;vn`C)}hbCH)QMAp^^X|e7jR8 z&(NYmcIP6x8j5^!BJ=b1j7CMv3{Qb5P~A+&AM+9)%(JGK_o3t+scgIEf z?e9uPdmM6gZy(CRfJ#Hlghu8M(k?%8-bVXJtg4m0XxZTZuJ1kM%lI+}`2tWSIRCau zI=%PUkt)!JOKcq|kNpbRpXedF^Ln^d_d5VZAs}^4UcsRuXTUAqn%ir;0{&4+kNoOR z(9BDe5F+Yt>A<>1W;|%N&+j9Wp?2VwcAkG5d+tX@A@HLl54ReHsLFjS;}TfuW(o7$ zmy^`L=g%sFUcRHFUn1vZdbR&M!fnFm#^5zt`?$dr^#3{Dc**vW(M;SgFF^6W+w+1H zo~gP%es5?Y=+%ERl(4Tz-l?IWNa!z&dHDn%?w|$@`uPI=fIEg(x*FmRrXJSol1B3P zq;!|jy=0?JYHN=q70w8Z1n1YijN&{N!p{qjR?F>o63=0JjVl8MKQSpFEp5jtIskxr z>TRzhUgU(>9mpbhf0Io+T6fYqFE z1K{v(Uh>;J^s%h)st*)N&&!;FR5Ggl(UCmIKy=&3ReTb*O95V77q*is7^V3@%Vx+_RR zg({lDRQ&rqe93{9DPmPGepUA}6hZjbd;szu@vWV(%c|Qt&h)ZDQhE%=aKHb3SKNX9 z@6`cgEsUG5(|4Qjd0Qh-w?@?zL8^*e;$l_eGPU2At{F8|y}q8*w~z|^dR#^>Wf5^D zy`$hr`Ad5*Du-vQw^bXW{EZ~%?_ceA(Ug@yT)}OZan|+B8VQgvt`i{NKspCFZgT+C z6P9LNBj)Cxnm_$$6pmR@HHb%^v7vBVib7xdbP*;+V9!cow-&UpUx-jn8tIaIL5NNR zVlWU9#Uhpf4=dH#NBnvYro(=hJapVIU;{s`sJs94Z&|#;Lghge0&xX{nFiJf8!33l zqmTKWJl0T%-9rI-W^25;NQD?~e0Zd8FaI>!hX_!B7kTE;jM6XQ=P9_?YZ=d=_dnBw zZcqpRCe0Drwa6LrY^$PCR_j^4G$qkQ(FuRlm#;4Um@NX*^T$ z6qPCAXBUD=s%CzK?d~q-lw`gB#@AkT9y59pZ+lY^dA*>2_r~7q(XjZkY{I~Su={YN z<03r|6c(S0c6HDiH81*+K5tuI-1BfJ6_W}=sTSW}3~?UlIB?IJ2g1cPr8_ib%Fg@NO^EW z6A35}G{FaiBJRy-a zLE3xaX`?%OdC=R3$IM@X`a10NAmPxE!-E9K<%;PH;vzj1d9`Xe9wM`b>u0hH_8Egj z;JcirEHZj{+@q3?n*B1s>7334#eWg0-};qw{Xo12Cif2)inSfZ*`F#iC-I}%db4$_ zmupKO=__Y0a8ru;15t5_{CSe?3(QfkMqYGHoh{$ib+hjfQ^IC#7MH=9S%#vC=`1w zZbdhG#u_hZUhQho1&8653TDaB3z#pjkNWex>Pk(kC`MJ6`mEbVqFSqvf)i39` zfF!3wLSS{TbuRrPkc-+xYO-r4zuLY%H~fyzef%To%(fKemLU8}pn!VGERzVxAa{r| zBlG0bcTl}fpY?awU|9-#@63$&!mzYpn;?N)A0_rIE(F|lcrgwZQn%V(ywMvtS4l3i zx-sKC*&oldJtTu%@U^cJrf~IlMBFQ4xs6fb`GdzmVR#)Z*W!J>3YP&;E(K!WBF2Uwl%>b0zS`Ae5&e7N|!oO5S8YWt;_(-6&l}# zfrWc$V=FAVw}1U^Jwe4U@TG7r7spF^VTPKu)`AT5y2Rpp8RJpQRW`?9NbHG1Q`|RjwL{#1{?2 zioN)O^s>2K3k|VT$YMVug}N;Ek24=JwM!UkEWGyg_y?EGog}|jB#He4ftJ&n#UE^wUdu$hL8f>#L!PcS({qVV?JgN!D+;~nN4PuT8h5#tJNYgn}z7G()U_AzfX=Xrn*xo z1?@%i5Fq%g{cHN5`EXNn{aIdz^P21z5|W6vHJ>a7gYBbpdvIT1wP9qhY#g{Hp?}Y9 z!NoCEo*kv102kg?=%lZ1)gA7o6Lees&57rGMSc$#mtpD&elIN`(|w^=c{8J7gO!a0 zMrZt3kl-~JlqU_;@o?`i9A+cHHq(AW59FnMpQwbwic4=?PE(~_Y8q1O9yDlARy=#E zq#;?0p?@#Qc?FdpPp%fLcf25Xtre~|Mq)~M{7@ZGUgyw8;8-nkLJo-Yaw_1m*I-4K zHzMYa%N=3RD(gtHCaKr0cVZkzS&=(}f`5TgDT99u(M8Vtvl7bG#0fXkKa zoAbO%F!&qhj}RYl`Zav4$Ce{xAI(kd1)8y=zB+{SLJAW>d-$xEvquxZrllwejW8%3&1*je0gf zC0)@D!qMopO-9sx(EI$B0%AgKOSbyNE~ZE%kqqd<(oVG3(xVkBI)&~;y&m7w;RWY4 zO+sljCFXoTn?Vt__639sV+3uzn;Z)m+u;ElP{fo!3(aza<8_zCT~Lk9&u8h`b|)o6 zvKM5HAU^H~g85GhTp7Vf{F#n6;xM-Gi_T5^(udE>6S924We$z_RrWVs1z3^ou=e27 z&jQCH?rZ)Sazfj11@@@KNQB(i>FHQOJfO#x1p|h~vWM&aacn7SlH-b4h!S8$os=8e zXFe~AmhIo`@GhW4K<>1M#$H({oNejE@Iec1KYQa3%lGT^>GH};LlhbtGe?>39ddOh z9=j;CtxS(w1}+(xhvVPq^hF=A>k}mf7Sg((_M`YIV+Jxh(%Biw@A&sRt#nUp_4p0- z+dwhpeQ!;@2djd8it};-X`>*nebKhCf?s~1@jWm}AOhUijPoMGdTuoN=$JDA#7|;D zp1RIYxyX^M1b_A3*gOKId(S0jpuZBow{b3?YJc_Kj)5PwAcec6;t!m=p%ki84o5xVfd80f$ppZsa3f&g3jw8S0aYw#d1Bp?&#*nvS}W$pHbwKg8%HEo2K zh&zpE8xxe`cZ0T}@p3bJM3lYG0wQ5I`Y zajkW<_h)O)^u6E?grN#z6{0XXr8j3hy~_D~P?u+X(zThvQ_V ze%G-BQ^6b%_BvI;hlO|%=TNB#Ds;J6;2AP!=@3#fzXz~^(H}%VNI?UQ-glH+@|h-8 z)KBfHg|ht-FP9NRvtyI_P3U>&D|3 z(YkC2(o%%H2El4k!iw%~R+q#6)g10ufQmO*L#s5@^A8GUCS*YW~ z&EG(51d-pzFf=7Q`MPEK<$gQGZ3MqQEnT3-sTJ%n0+8;!ryEGD+2M7osUHYhf&@^KD{gfE*@IQv*bN9%@Q>wh!&b* zS9ftg%pS7#VR;dvvRP=<-CxNx-K85qA)I1me~nq3xP*OT>Hz`}kdVMoxgK2`K0{Lk zthiLrE9j&i@&2lOA3RGNP#n1&GC`EZtISuqN}ClkJ0t7 z_q|zc=#2MaIw;QrNv3m5czL2b%6Aa2hA>sm)NG7fccJS9&S?KlMh7uhA% z4>g}aGl9l`U)&ZppBxIVD*joeWt$Jy=X>XS148vlnO>TZJR~Fh_3gEFK8E3mq1%4` z@2D-RFJG>$ zJ$ULq&yuvOz@mA??M}CodnxwfSY0avX{t}`>%xqYZE4W@#hBWL+_U58r%fDQ$Tw8) zpiGi1%BE~vQg1bov2KkoyvAI5j<}8*Yz2M66kT87#O^;H>bvaMV|dp;*w2%yE$E8D z1ZMFKFI$0{Z<%OFBAPC^2I9#Mu?zZHWlh=V{!=hy~CE0_B|Y^S^nxEnDKyk zFd-!ETuJ2o?vnvtOeto2-8ZX%i7osIbcIz$66(>#^{_y?XyVx!ddJ`q6}hJnH+nb>KBfV#a_- zI0$WU0)&}&0Q6{8TU@u!n{Xpd3@(7Ybu8=ZEN1aaqWY+Bix$+6VNzB-#DGFj&FgDgY*m)9t$8ltd@lV5(xa zJfj})GvH=t|J&li7qlM?!4{@MIDbLLq-eg5*7TJj=oEzI08xR@Qr%l@3*U)HlYUCG zZlnF}#x4?XZ8cSYT*&F66)*#E+C_4nigH#HNDkI07N0a>Q7MudpM@doZcFx$;7L8_ zVgrGYGsJg|2(uPhoJr96ePbg0OABEGcSyDmLg8M!wvVaC<@3D1J20aV!3kC&Hno-P zf0H&Nni%vq(SE`ZM=x!!3dWU$2I*eRsNn7xw-y64F+AGYz)Iq5-;yK}5R1D20K=Jf zwYs_<1`Roj6Rg)L(apJx zGE5f}dVN|kSdPPMEwqWCd@BRiOv~R;JN>--=d>bQfU)GyWJn6y_%0l&;LgJjjqLdI z;W@Q{mbF&u;s_5@y$L&v%5OLOU3BEdm)<7>%}_u-j8iAvD!)at&Jbmp6--(U>fxky z%NOBxjeSLg?Fpe6qh|UMvVb6TX|0eLn?uO#+s?N+oG9&=6SC2|D*E6U=-_TE?=2ce ztBsSr|2DbULPPTD;<1LS=aL78Ac^51j4w#X$k50&=i%9J7z5|F-H#pCrsj=y9(|Xq zG2m0iS1B|E2&D?`HT^=KBb|ZC?X@ z!Vn<0d=}}pV~GCEX;fFy&#&bJ7ILHoYugq2mm+p95f>n!T569?QQ{uA&kWHFukaPN^|rlJyqwTkM&ka z{BqYbf#5)oN8{WbV!$Gh^Cjdf8Vnu6b#dN@PAk{UF!?wX8s((*c1}k04#LsolIOWN zBfIBzyg+n!es0Cql@}VuJZQQ;mhf5su!kd>gYa`=uwZ@mhUqj4u;)SHi_?T#f?3?% zEQ4!@qYnNFKEIPh`{%O&dKx1|_(9$Wj7$BfN`xe)iq)ZgVLJ3ccX(Q$X0tA2 zH^yXVSEF)yku6yRHj+C;2T|oGe{#VWLPU4K2IfcWnE;|J&OffMDW777ek^NG&{X)` z{cFZs3drtF7y~4BWQdWb;Ctd~2=WwwWx^Ok4*bkIyVn+M)B{qokR-^J&Eg|xy@KKF zs$a1+aP|)w;GY@oczA1yKt<^sjNlUdLo`|BhC6Ggqxmdp=e@GAN6%k=gpv?mq)pgjd8 z8@c6op4%-7NAy%|jL3PlQgS%!A2r!GUA5l{M>c-#`wZ^}aH6L(l7*ZgD>?GsAC)P|%GzgXBy*l9+`y$CJ&k-T# z!+hzQ>BZUarmc(;B1`zUA-zqm7npvL{0>5+ltPeLJh(r-IG>=3D= ztWXbc?fjzw#JWNX$#Jm6YbH9BtV*B12w=CjQjN;Y?n-2*@HTzuFAvZwd_V2^p$N+? zcO1kmZ7-y-XU0VUF*M|0&ejLW18y2pd6>G+_AUoNA$z%lF01&Rjz2-``$l-?$UO|t zA{Ac^_)BW?odbY_FKatl&!NtE@K(GLQ#!ab8q=@K&%C~3qGN4D3~eKGJLn39ZZ>6G zDBHW!^>7>n=^ejCR{P=P28&zABGl+e&NK8RW11QJN&E-(2LQV7;s?gp_bdS0GOs+~ z_=x6A<4m|m&F-uyZ!Po0a?9Nc_vjWk_;7z|UmW0a_*}8f#mMyqw85Nny6^C1x(Rth z^r#(+o)IJlk6PxPq~Vzdh5Zx}KrkWHN3;7P+~UVwMxk!l_~c$@VgOrX z#x`187_|ic{70w6!R`y3;DCf01JXKs=?zT;`5F%s%pYIe@=5@xfwr#D;C)C2$*qjJ z8mfc9%T}O=R?@zB9hnLpRnxl`)(=xh9v!GtI6H?vcz6D?i{;QqD)pFkW;Rp`tt%6$(7LJ-v>vS0WEYjS!55qSFziMsyta#y#Q?QZ zsn_x9}CSkFXw7n?FQ&6#|`w5y7ejg&rz=8$Ci zd$ly!^VkNs_x-sbn2wzm0bRG|6C6zET;`8cVFG69=Ih>GFCLqV(2)XPk2g!7fGWCk zQi5nG9(ifbA7W}A&)S?4_fbk{s_l}aAb(GilF`lg{pcu#O`iK(WYMp+mjB>=D=riX z%x&94gOpFcPJ(Wcei^E76JHc3f0jN@E9v4x_I38JL9kU65{K3O?e^^f4(}X3jqAxl zWD*eb&a7JFA}FsG*kfXkoZJfsfPSB3$d_p}~AFAfKXqfHg%f2Zr7A&Xc&= zLd$iI&tYwNjw7gh-)FFbyOcX>eTDa=mC?Jb3!F=?U-EqYU9S+sd?jF?@C+xcTbp}k zJ{(Am`#lGn9iBlelb=w1-@``n!}$5GM^2k-i%shH&%QVh>|oXc$CQ#~d{M*mF?4_~ z)x;cJkXOS{psTEvJ6}&B#*jLpd1kb)^vU~yqYC%UJ5y^P-wY9y3mktpyKf>PTIeTU z0SwS~&_ZM+wYI+)TSLzEL#G)f<5@_1r;`lAcf*4HiO2!(-J>Gk`b#&aJbk>^KdKZ&xm+& z&ZH5A(`Od#7{u+=~as*UAl4mtzZ=RI+u zfVpFDYmkEpbOM|0N6-7qInql|qf)OdGf*tI6odJ^T<20w*!qRZ-PXNVEB`Tv4{CRb zAqd->qZ8g8r@J@L$>^@O)=fyY{WSt{ya%F<$k8`+PT;rwreuFE8&aXA5PvyS2n7Bi zI&}onqBKVsV7&EoISsUJgCB7mYUe|tH92ZlB+qY?qH)guM7r-?4UqtGsIFd!9ul^L zsl=x=m2y4*R@2S6{R;h6MZnN53^f-gg+?-*QUo8cQ^BHz9q@>Ic$F2PWoqP=?6Mq0 zv3m0hursusA?M&Vx%bGCTyQBObpG6n{OO7%XTPu~*n06Fe*)j-i1-hwb;hpKvbR|< zT*sQn0?pGy@Ia#8(`_V_LFwLAU3hSj*!w8%OAxYMCIqu0cJQ4nHp1p?MEoIDybD4z zyB5C~VG_`V#|r-6L%z`7S9c|Bu{$*l1r_}9at-Wt&Li$_J%=Z+WsugVnx)UKPsjQ1 zoTVRldl!^eP76cwg?lkfd8C0A!>k@+(5bZ{9O+;Md3Q%rMeFo$oH8Kbb99Jt904?*l_t`FdV893@2SulZtvr5V6If4r`Me%WUR12A~#ONRTtG4uzQ-0riR#pe-$+;g0 zd}mfQHNt;_n*fR1>(bt2nKYA>XR#PaDd{i|@3#C+fldb#CgI|Z1}UR2Bqph1z+G8a zs-e)*IREkVl@E2o$@1m@Sa!_q1?;C%+tNbgK?bE|ILJ}XTv%vqJC|+cn1(MOnE?6- zG1T{r%r-13$`862MS;o*qT>Diy`44JJABfWR$2vNb{%KMtAI%9C`CbIGq3!(QMtAu zf7u1Z*4TQH?yG_#=?j+|h~KrA6G`WQMj`DC#y-f0*d8E#bmo}o^xapUnlyftR_=vy zT>A7oi8Sz}yBn5zsfcOYN(=OH34LD*q=}uxRz1NEC%KQRO_?i90KU#Exm)rN_sVzJXAgV!oXa~(isd7oQqq@^m=_d(pi?GV)c zzdr}?LYvB)_S(gii+ee>zh^`ea!_Ey=WzrBOw-C&MbfFwjUll9~BRedlh`9?f) zt;?+-9utTe;?V1{4^f4oZMaJ{#CpN*+#R)M#IxDgfin(zaziG;E5W~dC+z~+9p+

vg>{_hBLrB(5ByCPh=`wmM6o(m~|#&V_%Fgnj~>Tyk-vvY$k z#7FXEEu%jK+q#B(SMqg{UXb0*QQ@X=s?tR{IKZbFgv|mc4;+i1eEROa+a2kHg=Bv< zu_Qjh`7kT>;kUqX`!iwO7B#J7iF>fDjH;DtRYIDxl5=b(#<=XSApz5~$$6tUel%uXAk=)nY zmmyGG0;%pR;*g9aN9=>O_G~rs@;!de{2Rh_CSN>EAFS{u)B+!T5f4Dj$A~G8ZF<}n+k#OD8!g<*@(7`2*;oCZ_m%;r5X4b&AS6`* zTD1J)(vg^hYux#Y^#@jli? zi%1nW)$;2GGVR|`LvH%;$r}6YbwL6Ze8hM*RC|dun35jEpdY2rdl1Z{Cs&@#v&t4_ zx8KHeMpI3Fk$p_E;yR+U!*8!Xb;->#&XRM)Ks*W4r6*Z^pod54XBxd=8L+D zc?(hrkN{v^!}{BIr7Kb;m-z#S6WC<^I-6cB>eW~+1P)9S!)_{L9Z2GX?H z3w$66lot+XdHl<}+C_p$M?@V-Fk<7YD(98t$KfFBxT=6kq2B>>u>$fu_!p!w`jD|G zkCM~NeRFy@?&(H-n(J*tC_*h84G+x}BA%ez7UO|ec|a+U=G>L;9OELv#dAW!Sgx_a zS514vA)|rk><5IiY~#@e%o27wQFXL#Pw3!iRYA<#(f%G+QPlk5eQ~X?C-4yLZ;>D4 zZ1U{!s2xZv;?Nw4SHB&eC-Nl@CK91E4-766ijn2#;=xY$LSK8t5&SM{X?)yYVr1IG zO}+7*z9wHn@E?mf)bbfoR@Ni&MrDyDXX#vh)xWKi3e`UR5jl`-Rlz(l(e6q^i04rm zn7d&;skdfylG=Y7QnDYPu{(*>TXvh1CEI|chWj%;-?t57Fo5475m1#!cX=J|dcBxB z;&OlpPdWdk!bG^9ClB!E$}rlug&~{M$Mcl{Zb$uiO{-ORM02IX}qGwng2; zfGn@jh&-=74ZIY<%&C}UDQ_S1z)lAPD6%~GT%<$qOdDgfiM%-z(82(RK zBVGuiKOcZcO$N7?neM7xd%(c~0zZEqbKzY#s5pK#29yuPqiHY!ppD#ZES=6q@&ZsC z3?DxIDsZj@r(Q5_mrFNObu9zX3YeO@^B%*BF9Je39OQ|tGRY0*Hx=d=93eq1jOWw9 zB%}|;+1g5#mC=j+1>1Y;@8h_UKn?2`Nw4R~f_4}$08aBMc7180>Ic9{sna1(EGh?x z`7bjD^hiNJ7WVs$orsai3xV?hyi_B8cOyi=Xx_x8<1H*WnH%2SQh0?BTI|O-;iEZo zxJpfG-!B4Vhb+ud7opP%VBp0w(~57_JDF%}ab^rsoEM+y{WgtoQeO z`k`g}LC~X$2_sPvA+48KtgfNsr?FlDaWu&bb?R}3O_;;pr};|VeVvfmpg+V)+^E;E zeUXI#vHvTsPGB;v`_tzgxLjdD*u=@fLwpd(rP4_vWcRg~G@ z`TlJNS@D0lW-x$hffjcy-_cQv#mn9DO_BIU_xs-9SPH>N?bPFfxB9$Y#yiB-f#_j` zHEI1VrYA;tY{Ug9<~^HFq8e%ylFQv(Q1>7+iNVn2qLU!JgeC~BtE~Z|Vp^x?Ud}{U z5Rx?VR$RVIpaYBU&$Z1w)HZrM$i7nx9>6W^f}A3pxQsL}dcfJ@*_5rSdE6{)FR3Is zuvo|ljq;d}y=FK%15?{!I=y1!hd>yP2@-DCYcfMTe;K?o_N-1;qH^!Dd-dhq6C~t9 z{hw#IV)##ZL;^B#TD5xx?Y@I}mm@$@M7r-slD_ZXcw9oE#0vuGzF!`B(ZZgO>UDeJ zgK2S`7@ja=oz${XtJ}-iU2>nQTYfzp-*~N$WE}Z@gdl_?1wQuQ`({BKYanHQE9u`< zPlv;aWg;OF!G2T$7B2-&#HMiTKI6V?>m;qcK_eAAFavG`U4Kq^PJ84hhXIZV=?{p< z-FA1~<59AIO4Rnij#X>et}FePpM~XtP38cDVb*xK`9XF#VOfxs^N>S%mm>WE%+rK~ zXr2a?_4`QxE$I`9WZc{W{jb-=EUSL}s^21#BS*X5Pg&CKfnEttK&waPrvLmveBXHk2L#IFNrT9$K`?qPHi>hteTVjw?U5QL-$xZZ;f zidsBaOPxMGqFOh@O3ywL7E}w%hSyh)-kiRILy=^oZ7H(T=Z2^qaV07@#@wxyjnE zf{tp_R@=pM3v&zb0Lybq-d|b5(cWDs3LkmthX43w*Kk58-(jYX_%>Jy)6D{VMH1A^ zf?*mW%ML_X6R#6@wZew8GuwkP`%@72BI%NVE*AGYA{L>#eHh3pTgHvPUrVa)H=;I^ z0-fVUK~xvfx)^#`knEFs#Nc9pch5K|c!wmLM2a|kG2Ra9KnavXZK`sdV|nIc$=7U- zeo-*d@>*pEMAZ*?CWUa>Pt##Plk2N&(?pL|R0Ju`9QGMdkM5ErZN^CH`!KmKausN~ znv#nwM=oTX`YnZ`=kW@#y|l9Z1ipKydc*WI{zlisaX^C9HcylFm9CM+}UBaY$L`G`30j@e2VcPg0fzk_pZNFfHj!AD5nDu z20L#p+aLuHfJNqMq0X-A3O1<&6G{~sBK&k{ z-GEQgk~cXh>0jUs^A@a(IN}QOoPayDQjU0}SLXQl|$O$`9z#~ik$$Oww1VJx9U;#@^?S}vOILA-|IjTZQ^dTsUU7d z92AW^!V~Z_Zg)7Kf4MG+dBP_Y4eY34KIutS3c^~eF z6o`kHdB5iIV7|{=k|NO|4Y&3&IA6aCB2x3{7j7vHlHdwBJ!$^TZ4UTh^$M-a12Om0 z@xCAf(k@rK51W;4)p|@#+$6*Mo-<@9!H&m zxm}T6wDMG;yl?Xc=gl+D;Bc!&=>c6Il#!>2-0v7>l4Zjfq!;?$_fLn+NNd4t$2nE` zmSg;Oae%1obA$F2&4W+KJCC-5os4^2Yl;y>HdlSw%8onW7R}0|w!#f$P=$I%QsX_D z4RZ|48VcMD8Gh+ptCw_N#QO$w4!aD+TZHcZt>dx%`PKWXc77I1U<>Z&TLy|hdj8^I zQr0lD;v?O72Av*`2AZhy(}481S5g5=N=@rWS(LctPmS+3Dmsa`IXQNpDxYHU5C<6Q z8Uag`WSTkVy1*(b4`1 zsQ9+{jqIlDK~#3F!g~tB+PvTUs9F%Y)gNz{oWNa+}Gt2Zbc_84NA|ePtULL zz8DmCVc@JPGnlgRHSFUqwC4Tkb`vU!OAAnht;ex5%Q*m{&F~7av@ZtP5uw4OD3oap-=E4#G+tm%2$9%5!}=jYPo<~bc=><0=4 zh~)Z!uOni%-c1aD_`NRg_bFoUps~KQ5u*{l={0x?Wlmt&x3e_5p%JooWYQ(=^_NB| z-euO%fbO$wzm2}e1%2b;NJ;V-c2u;*ndY#9rMd6R?IewF9d04q7Lesm@42f@@@ zcMC8PB00Us_k+R!n!;L{=w=yXYRceL&gfq)S4Lh+5x z&l!Hxq)mg&Vsy8R{Oe7<|3s4o8VR4-lQc2I5IC(u+1Am=@R&4rxU!d`dMhCaxae>- ziBqoxq1F|t_aZW|wz5KF@h6Xaxr$v5B}M$SO+6pwLMq(p=V)jNG*#;c zr3o@f1vKpR^L%ZY8w6Oo9`LYvpj$28PsfX|bau+*;OshNz(J0E>yEd%|Cx|1e@(PA z5aEEGLDE3sIs9&km~d3dj8_KI)mICKdr6pd{5b(bOJrx=M#^Rv6)L9WWme1{4R;~D z|F65wIe{fWKXN43>uqI4A~V^D(=hit3Bt81N%ze+&u4r74$^*!V>H6a4dwy}&$YN? zrJ2hZ3;uM92j5=489*)giaI6xb0#ZQe~s=301u8xtnY*Mu0P^IF$qfx_nU~l_2*}R znrndjE`)xBHvheer1X-|oG0>p4l~W9n#ja1sSvMZG=S@sEiHu9JW{ zjXZ4+bz5AFsoCFg%TIF%hSTnbn5ftWC((HtEE0+!PvKH}3=(I;V^RrdM>Mo&P!r-u z8e5ofKJd*|bb}wut!T_^c|^z`f4U#Iq=v7I`nOQ3F`@wA+*YNedh)#liv%e66|{?J z+?Z!iB3d(Q2yhITx8W-Y4(-@%fYJQ(Jxnn3k^L~|?HiH<(dxDiM1E)$R%JT$eCh*CpBeyv~!?aVM_FMlP zWtJ}1qt7u6nba}p%i0gV)Li|Am@?v34xCG0Bt6j9SI}Jc#HdLVKQeNlF78zpeUQ#} zd~usE@fj3fN383*NcK9=c(U=nicq=hNIScT9>DfjD4@InQy_KZ) zX#LZ%CC|xUD}l1hN&MNV1{Dos)_52QZH`lVq{rep8!lwY!`83UgJAAAHG1ZO&nE4dZ=9RNG014ZZ6P&vPJ6n5`g~GgsdjJvM**ITHEQiKD zJl=)ZKg5CR!283-;hBZ_8bGFhlE2RbodGUf0kJxxH64E;;t(^KuuGd|uklH6!KbW? z{l0DnwAMUZv0C2KbmTwdV+u0}bX@G4Ul?hauIHfyyr|1ulB+$wqw*=`lHKIRV*)w! z{5Y{BI-rl}Q@_F`2T3OO_;e^I;N1E3oY91-_@^}<-tO1EyuZi8?d;P4XueS=G>r4G z?)z^AklBgCcq(2N7WLv@#0t?Hn!J^Gngl|$B&BevXWC7Ng=SNqNva%UpSbT{v5119 zWlnc1+xEK~9>ZT%1GRfzj^B$Gow0qSEnFDr8YMmPWa5hrXx{IpEvIFAmznV3zpAYf z<$%x}q))ou`$*sE0@4J;n%mjF6TlwX5YkfxGEIpxh^FJ`qA3~rIH3g!-AvHXvF@KTt9P|NUgK&7Ia-=Pz?@T{@NeZLokO?ESePW zuC-i%!xG$L5-rG|Dj^ke{Pm3gjPzA~t1xf9H72!b+50?T;DUvNi#sqWZecvY$e?)TjuBtz`alCR*jfm2 z>)a3Se-5Dy`=xn5?ALGTv}NZ>UpIK@t7FY1F9R-+Eb{VC$;jT)9cTa?&*cI6rUx#nl1ddN`sPf!MV$z9}ao zds=TXealFI{XnV$fwJ%E4mdLNr*5H6K&c9L3_7%WMNC^NUsKYi{JEj{&~9$diyx+X z%W1Wv$i0I&&n>y2tirzRuw-XAmq&XdmqA0m0_&V!-cuk=VEc2VK7`KVTRFr8mXN@N z^t)Vt@hfa-R(n@i?CiKjEScL7^yK$J+lO5Nd@lV#04k?j*0fcbQ%?K+mdgX_{Eo57 z;|)**$>%Dsc4N{yV2(oQM2DbU8?spkNv{{ogOJVxDn-a({UX+c7bk)X^-QVC^j12M zZUG4)1Of9x1Q>w#s=i*I0j8uO8mp_f-H!Sdx+W+xl~O-#;^+lP7ifA+n1re~F-E4p z;a7S^&=ioM)$(enK0d$pqax#~ZvX(obO>JU@mRSWJ3{bMM$FyLkrFptS=jHP2z=sI zN8-{;uwmpxRpvUFaKl+vdisp4Ilxp`In1{5%Xwna50ILV7zK9NSIghy8$TQU4BJR{ z-kBkMWi(lRH@@rLy9|#!WcjSoM5vqT#X~)EC($0@PV*4ozDsefj}2Ae5x%SXtD$PDnA^2PiF#^1q_zQYG43 zc3%gOFV&V<2)}oGncgq}-xI-Zt@P zuH6qBQOJRR@dy$@Gl{DY`x*(iiL(b=NzfiDY8*%$f)&o`iA?N$wNS%X=emDi_q}lP zV>u^ri1eEp3#qGvgufEk`{ANJgG4r&g1 zj8BHkR`QAf_tdIFrDA&z4t{~^Mr6!&zi4>$#PHkwpm(q69eX5Q;MxdoK&DEcSzp)M z+#gst;w|N_D4p~b6pCb2o(OL4NZ{h@+eh;S{gEjT(~PdjxcmipBH+3C;0nb9Uek#&^e z8+Z}sI+<~PUKk(Pp_E*E##1&Y3UA9bFM}+BVyQ0E>@9Xa+N8IFI-d1=K{}j~hnX5e z5Bdi08!;t82bJxR+$Nmz!Z{=0r55Bucf1gJ3NBm}ha)!wX}WJm0-_ZeNo60g34k1u zYFZzzMDdx5U_+lel8MI;Uo}`KqRp#U`*U2#&ys8^|}ZjsI*C zRr9Qrh=Rd?0~W$!Kh14K{}EkHgt#9;?l$>Jy{2@ThF{oEVWx4%0HDpMjTqkh*+Sw2 zv}X^Ie7&CTc-mLPqlA!&tkAD!+tt^WmqB{bZpcU{&)C{piiZJUd8L}Pe8?J1u!Tt0 zzcnGHC+oP3N{cbhXsOWhKt2_hdp}J|gqI9KhTc}8XGr^&UxG2sNxgfo=(kQ81&ud= zys9MlqUUFW$*EdV(cUp}=(s$+O)$p*(OGS=SQaaUC(RHIM;so%WWN9dz!#{K7_59` zbJ?3(h{^#$fc6k7+!o?v@MV?3E>k@e)9SZLkH=&rI?3^ggQFo{FBAFSb7HcW zlb!wck;Dg7Adx6_k6KFV?pOnn9=slXo;ENq14n4N7Zwh=`=d>&*+H9XV+28Y^1p+_ ze){o=U?}{Zh!o$)lO%|KZ!j!unuYs?j2mJ_bNjVjQUTb(lCoYy-UO!(qiFAp&)Tuf zH1!o*dbbz4V4THqeSiEUo&r_$D!ZF#Reuq5-&CHTAHONn^9BWqRFy8Nib!@)fw zM9BkUKWnnAcEtJl$E(~P4{pEC`gVN+E(adf6C(S2v(KF9?>Dta*~IB*HH7`gY99hre3x5kFI=1r*wa@dL(%xgND1M!G6}f?q^QCpBd-=Wrg5FtVkx& zx5^Lb)PxLXcJRZvN`MmfnaaOYp6=T!vsZ2iIL<&}H0UaTD%2VuErCzn@A6V`E^msw zzK_?o^ZJyadzlisS!XHz+b?wwTNhvf0H+}AiJM&n(lKO58u}T*yob6Q?!Tj;7j9z` z=|w$zzOu-%Ap)A+o9uGFUNA$t=TLG#B}L`ugWw+S#N~h)BY zq>Jnjq|;u$2g9ljqbnyz1gGGS9DQa@MIkoOs=Dcn{fx7M*O8&FYeR(l&}`j?XIY|z z7laa?W-4ax*<(LaVj{V5GO8r6LjH(0k4`qcOprVe;t$mYV406u=#fUk<^or6whIKq z{&?vHh(`zK>C<<@!~{x}BE%dhb&O_LGViOlh3hPb8+31zhjvgN zrr$@FV^9SqeMA01QE6TKoR0J9MEzFq>$FW%goGh4FH*nl5?$*tBci9#6-Sp)3(zbx z(r+PG$8-Z$jFokS7c@EeagM8HrQ}$jGd98K*fxe6cF;iZD z6rX}V5fRGtd|>62OmG13VDs`?-#;us9%mQO9$psmJCK>>hb!y0Sqn5;h*gxM>#HHG zZ-ylZQ@svk(De3(<}Egb)l!?saQre9TEDgq+%1~If(s~;ojZpzNEV3l@alX2&1RW1TRIo+n&MISFxId0)OsY+A-0ddB2_AJmwr z?H$*4|Iu_^QMLWvw~zxM<{=a`M0bxAMcJ}LZzE0AJ98G)cMnXZ@I}DUu};TnHtV3m zYOw0N<9pF%h>^qkwiLNI4{87xH5PhMo;Rc_l2gvei|Y%tTlZgDAvreEEwJ7F7%1v_u;h3DUNL%pi>CsU2|na3_DxFG5bIIA?KM z&^R`A?Ks*aV!lRJ*}aO1ws*dO36m4Lx*B_Nu-(h_W!zswaFbRjCftaJE0Va^tnx0g zlPn+T8D4L}P?&gcbcEUm0cH4$ic>h((g3otY%yKE^V-1yW?y=#o8#&AtN=WG!A;yM zOkVBFAEHPe=J`Gl$vv<_!vhl5aR>XM`N#J`cZc6djM)l8t_3<}y!6n!@+1R{yJu_; zi<{2gYA$~0S*Vw1sG>N6MWzP!Pxeu1=7O_PrdwEm4x!~Wj+OL9cdDQZ%zl-%+0v7cHK zjmYj+0i7EBJ5M{X?LN2x;}2dT`&1LO^gXCZSJ7L9d-UlyvN_~C8S)^dId^$26ddTV z`~hJiAUhR}4T;?MUn8^9p24TY9kvm!u;gfkDDX*LB(<#7~cC_`7`; z&ae2z;=~3!e9SMpNZN&wLX=+KuP+WWGtj4>KUKBF#@N5_egz=VNShT2R3|T#j@)Bn zpV9G3ELDu1VI6}>>0aw~k+$&)XvKN)rM&-R3HIpFX_h6=d#8r`oAt3T3nqus_Zr86 zqQYG4DTSQsd7G-odJ~rom$mM}A_z`uoGGxK?GxMdV_%!e@8s;w!#2S{9pd0e5Yoe_-V%Lb&m z^~E#Tll8gH`^JE=xzQqB2_c(m)geF&#wvq+j+Xf_o7j6QFz;iU$CH^X!(> zV0Z$3e~CvIiRR#lEm`SFN?<(^dnx^TBz#uZb<>V}(eX){Ct_VfDyXz&3E)K2e@+eh z-GO_qm~%TH=PjLWHQjIIn$zA#M+ldgRG}};j*D3`QSKZDq`Zkx5-&+`IE5*>9#E~n zDmcc4@OUsW5BQ1v@&YOPD#=fukeLrm2t;;6H0|HMOw?$rf{7NOboxRW4JY{kC!W0^ z8Z`R?##PbjK6BGSwnBgp9AsR)cPg&hCk(SkfJzm`80X~q{e+?A6^;HM{+|cm;05Z6 zO!9$iz7?}aN+27`GolDP{}FZzl}>!CV;~h*SF|@3M(B&plBVY6D|L}LYCDoiSVh&%RWqnqczd?X|}s-q~sBG;pz zgI;w$ZJ~F!%L`z|gC7{sr7{d#!UA=}cngUFGc})iL~wA4{f3!${hG2c zQG$6r?6>|C3Kg;(CYXMaJth)A@uSTC>|ORR0&9X(1WUlxT^8pBE8L3D04o=F#N76` zi{l%wo$tjoTF^q39J&DjP_gJZMyHHiGV(1Z#7VT14QTB-?$3Mw>rmziPQTYj^pdXB z`a)75B8HDg#P@-2b$Y%SYB+m5ZWHO}J>h}tL<5HtZqHrSpPh3Gz7J+Uld4{toE_~~ z6cClf=V*MJuDeEWxe$moJqIOstRd%SX$i4j50Iho82n0BuQ~ia*lgh3iuKTSJRzvy zDPz6#;-U8}yYrXS#W)8Zfv2OZ$oUKm4a)KJtU9EVEjW8Da4F~?3I{%xo8vh*4q2?` z`;IqKH6kOtR<9eRA@!;O1v(OO4$}jn5{NW|u%2%6{19Mr$Wl`Di4m$ptbLA`7PW#- zK-ub-y?aF>{V0FZ!$IJX@?i14S30ifPn6$0Llv6#bw3VJdwb^$9)ZaT6j@xwV8I(Y zv0VH3PIzuwc$C}ediCYoRrz(>5G|g9%*GdUk+nN}x={;;erz1KW@?prc3b7TsPnHf+2ua3x5mr&26{tN{WJ*~FQ`X)NT$Ra`EUY^dMXNNB!b^5cRl#_<7(;t1a0iJmWtwZSBVc%ccCzQ z{erzjxcKD>1U=B?eTAp)FO{YBXa051aKfyh0d}UtzLATua2YyXPed4G7EDVimK4s@ zKk0D|zI*TpLJ{0;Kfl$-QXNk(ns;Kn^ZRT0oznJ#3lU$$tzDF|D@5o4doB(OmkJPh zO_o5maMk;z@2oLm(hp!=pJ|09SxH^*(`+JCPfYnjb$dF<<{Ym#>yFx#R`ymv0<2@| z$_2thk?aZJm0I=YTeJ&{=rFC0@?E$@)|(OQQvj&4x~eh1}mz1bM%xdT~K?fkN+Xoga+3(ArW3XS5@vXA##jyv~3F|LTS@7H2T|pn{{4?JOBelS#%3Y+vI2 z>5(qhe>xp{JK(^yyGeK==cMcxJH4c+)LK#>^y@kx+phoIJp%6>62~vrbfe-|%E zIDZ@B!G{Rxw2IjRuZZ~O9`iXG4%ulwc0bO7mmzITH}lmJ;@yQk$RxfG&L+%f2D-+L zz{$m5uz#=sL(s$+#6SLBKUxa@-!)6WF_R~4%&@>2){F~qxr}j<;M`ytcvzo1e+3VZ zXD8JtTuHxo%l&X%R()wznEYiR(cau;iWmzNPBx<-mAd4U>?$}kP&`fF1Om^#UK`Z7 zcmvPmF`M8qeep4prtp>DV=&nmU1uJSAPhI|5!se}^M-7-!2v{!)Jt`s5k(#861yBE z|0v0O@p7iBtR2-xy1D^m6@2JG3C|XeSR6MkX)k(&&^2PcH!8zR^7hL#(}~grI;7_S zdv$^lzj}6!0l^G-6)#`Z3FKl#{a$pN)bYa$_xkYQ4#;hh3e+dkz9`dU)wPWu?I)5~ zqPjGX=HGaGMzIiQv?TdEjr(uOKd$@6wg?4I6=&v4V)G4#61_Tlx^JY+68f#Z>1vI3 zDs}UbcXYCF3vlu=tP*CwkPU4wh%rqJmJZOi2;$H&%wi+#munZW2^xsH_xL(n3@(OO z-coD|P(qP@IYLJAjydPkTpu?ra4+}jcAl4ItX&bK+l=#`4B}moQuCeytZ9eUQk_Yy z92GZwlPiDuq~TYn&3*5X%1nP0JXgDd!xIkD-*WCG67y~c=fo{D?HcVjjO+aEmVTfT z=}T{uz4iLEn`x0xuh-Z4eHtj-!r&|M6<>z<=j+VZ2V_-s#894b3P3K~kB{o%9r~4- zw&WiUa^9aIcULw-=#aw)wINwAy>rLz4xp;;AJm6_CC)C5MGt9RFX2N0;6ngzE?yc1 z=&K09kz|%W#Vqpjrc&Zl5He(6n?%V;>l$WF;FzzZx!~Cj#`%lwW$3y{66>>vW()10C zW^~p84#(Hf0VQcW1zHkWt#z@w?BjFfcO z3Qd0x@A2?>Q4&|V)M-6leDl>nDJ1RR>y;>#pB=7-X~{Pi2zBcvw@2n;Sq$()=(KPO zimv{rv2lpK?0_dgub8`2M=adTZZBu_LJilr^ak=SC}uqtJ$9r^WZ@~-Nqa{4=Me5G zLaZ`s)0gF**9MJz7x;aoW##`Yu1CO(>EIIC&wEV7z7?C9O1VtnrJ*lqEhL-3<{vTS z=G#U20?!doy?2kCe zoF|HA?rX4D|Hj&9g~i*T&(mn*Me`=0cW+G> zX+*K56Ci5$5Nvew1uL|!XYau{Sph3G#jh{UVji{s%euZiQ>=cayY){^OHM0WgI)p&pzmH z;KT8pgHg61eo8aeL{6Z=#=st7lBGDL2(+l}J{QU0VTfs$6-@ZlUzpU*%62_rNa^zi z+HAUE;Wwl)J*It=Vzp#OIMNWNA2|Sw?IXKSU4NA60d>#Nqd%zK= zOuyqr9!)TVkayt*Hh5N2-Yg%vQ4W{G8h^uX_6`sQ2kN{39dvIhj7yC4ZQ+=+Zfi_u zfx&Z!vq&NXPo8B#JV#gw5c&)=J$1(f5N)B5%@L;V;&?7JT9z`MaE{maEWH2~6GIbf z`YkL@rAUV1;}?*itiUAxsBeghXT(yYeDo9@3D?UK?eL_Eo2bkpcuPDC6WE>?4KQ_~HgmwMsFe1PA zcXoY{Ytb<7I-T}I|5+}t`EB30Mz?z^etd^Cc57hjSYKW2uZRVaZ%?_^KcICL7TTLt zurDvatbEC*)}e0fzx#*2=?^(O#Xv`0ePX#FLyQYV)wH%J6N~d?)FIi8QM!ebumLKPB zo>oV3rd>zz@Yxck*}BOLWbvgZ3$i}Nulg1U;Z;a;LJ5U6{aE`c|xrb7B>@x-b9!WFx3kh#t z;R?zsUB;ME*64=Kl_Qq@!Y^x!3YI%ycb=)ZQHygi9KK1Ti(Tx}J5!?%LVe{c2cWt;9qmt}5ZvM+j_w_a8WII9 z`LfrY_|?QeQ}(<0Cq5;<+(-T6{DU_7m&37QGNn3ax=X^?UmQERMX{h<9;Y+C#`k9t zHAFhcZQz}fR5eYb+`~xvtR3FbJ^`qqEC-c?ep(vbzDcHuE{-T!jURmf;n2(G1sf%P zvcW5^20Sk`(Ui91((7{yhYVM)h%J;ooS*Q|#an49+5;Ens@m@~C!k5irXbmt=%)|B ze6B{h0ti13xA5q+xu)4Up#f(8Qw50C7QAViH{M$R*q3?4+P{x;e?d?<_^{uon?~Lm zc;#>aWYo8Z`~#%cuqvG&PUN>Tj387ju<4}7125&-3J7lfI>F8UMSi@qQYcV-4Xeez zLLKtvv|l%J4(I55zH1nC`TZc!rrG-@FK2+6D+7S`x;1fdvBD#hIdNe z+1v^92647`jub%WdggZr13g5iMT+n%QWn&N^pI8qb!QlFC=ge!_gjG+oE=WbUHxY( zeFg5`PdEBH6oq)?>qJ#~d{@lF^fzO@h3zl&nPtv0T2{@PcjgSAGJf{?@Cn@g0|CyIQHX-rd`|TIX<8}2e zNEr~I6T1P%<)+mkAuo^UftK_)EB1U8BkGV}ZzB6sw8pYTWqqME(&QD>M&A;xiQwDpP2(!w+Y2xiTL zBUN8YkDD(qyxH8`>e)`7mY}RCak+rJ1{@dGr@GF5KbA~q5_2A;IyK+AfNLu^>8nU2RA)lr#pV0>Brlkshxbjp&S!BBHp`CQ7+8#fd*ji+7M_N872bEixe1q&N>qN z&fcIoAo)J~#_ky>`(>t9Pcm)ewlS3SF-tCw1fDBI5%dLQ>4ItxIyAt*KurM`-Qdny zGK)B}aE^)_&{TWXgI@~*s4^);6rg8z@$&qt(4SK`@`trQXUkp9UHW)&_&mSUEo9@zLQhPZTfcpUTwBkKOTOC<6^Jj|qfHN3Ho%W&^&N^bDYwhZ};Pj5;Lu z6EVXMKe=*<;Y-$muA+XRjr`~a3fnJ{(9j^(qK&g$=EW%%&|7~m><7U5)%)-LQ2&5C zdueRA((ijJuQIZ8yiyVk`pcfo<@F@B8=k(byLE(+yQ82Kt?h0~P<=Vl8j-n>A-Y{- zbf^~MB4&N_(eZ6hxC{~L4S7`~7Iufj_hb9UT7b^MBkH!_-$apa#4ln=^7jwObJlZR zNRaVTTIwCuAEn$vtRrl0f*=p>b=Km8qm#EER5Ep0r$8y{UUI`3nZ|x8e;Qg0U3I?mSaG=P?NiV}JwXx^ z0;hFUi|)jJ^a|7B2J#HEnpp_um)q~-?Y-!Z`IW;Bd@riw!v-S?kQ%QPNmJjUtq+N9 zOi-gn$1l`bh;ohU6xuD1Tq?38z+c$H2h!USve3>5OWf&-wTQq%yTI~4S7AahSPEQU zJP-B_N0B@nUo@H7M)Ow7-kTA1JH4M@EONrfOOdhD4edc5 zwie#uqrkfdqd)-?fJb62)skZ$w#*A5B=`%wlKHkbo>8WudX=wyLyRzVI8M&Fk2Cvt zz2DF8_+2UnE03^r74ekz_fnFaMbt_YT{6>St-oV-y;xUs-wBHSs{MUNap@W9nG6}H z&v#%?zTt|K%YAn7%og36d<`;Ot_in9kI!3l#QHs7e^pqw^y;HoBboGlc?;U;>`i63 zEU;a{8mOS1aT6gu6X3bS7bsBA)9be~m#2*s4kWhgkXa%9rtDEYu-*^8%mV}hX}I9X z#cAL2v(ktTbpy*010C3eW{NC79ZsZduGWtc7tugqF+3kKK%V@C16)i3V)(@;wpp$I}HV6$1 zp&j#}SnaRj9H2j9=bLxH5p3>R-TA>(JX|_radSp54J@5O|Jd&JH3L=?4~h5jC>JZ! z_``(G8#nJ!Iyi6+kr--xDvAR15qMU#8Im-~JpBoTCW55`J0cl-J}_dy=(zH=D0-Is z{T@kHLBL@DHX*3*J;sG6Jp1TVAZ&{DMV`3#B=zzTs^UB3SqS)4x9wwsq?P*_nk?9j z^ud%I^?N*xw`qaBEIMf415!w|Z3u7zRt^X9I+1A-1n*3;aq7~Nnn;@o^$0dlPL~<7 z#M;)geAvUvOpk0hoz>yY2FYAsHL2)@&qf;n zTr7AXulM(4Lmts{+jK#HKE3s#$p#&b4%GFG_m8#UDKjN_wkN$2*%s8Dj_!Pae83oN z_4!%^e4#!&Xq1>s7T(6F0u;74r9r-F^Lpxf?nFVOcZOx)n=~F`L0&Uz}N_XjfF-(zFtoHv~`sN1}J!Hm@;7Ar9WJ>Q%^Z_0&dgDl48{1 z{_x-X`HV-R-{0NV;hb5TOju?36f;=@YX~@kx3-ZP90bpm05=?p_NsYS={lQ>9Bt% z*v5`(*&@Q#oYeL9FbT;8P;#ZxY>GYro#_Njm1rppz&K=2ENxMk!nQ2{WdfF(r#xqL z`|jO9#?67bik`r2dXc{aI+J`n<7DA&JRit+S*;`99_qw?FNiupsN(|$*3*1(Fd&Yh zYh0R{<30uRwsx+Hi7yXV_rvkS)8estV4G$SxnLBcH$MIn|XDG5l(Jf!Gq zt_V65OIH%?e)e=bhUJ72INL$_B*j0^LmICEaEpMr0QTz6kGWOOWjUtgU3V#Nd4`k; z>Fp{uaB|&c_0%KfmX8y~NQbq;)9M)~%8wiwUZADB@bw$ngP2#2i3!=VtPC~-G=4ZC z*xb+ymNQ|0_aPmAvGu_+#DUD_0Uq2>85pYW8Qtkh@n_1%i81l(uM;ApCbJvxe1Nh2 zJXCOFkWcY3g5yX_ia@=8*@#?u@#cL_0Y4{^grN^yY?HERFqmeb-%4(sgD6UeEM0w_ z&Xz^ih*)C!EJ8XUY%=lxZ_MlE_zuyr3>>jP$v*b>yTZ1d;MY@ z#|QreE!rQ_vyMrE?cSR7gGNAbh1@RO< za^_1#f!?Oqf3^Z{$YrAW{bZ+BneHe9m$!4?;ITar)eX~3g=jP(8WK%h!| z)X8V)`7U_u>$^Tjq)~i)2ZnTwDOrdS+XZgp6~usg`s|x7tV!7dCLbBd)Ey5>f7fzI4hKC&j)_xHj9M3pgzZqDqSkuK(Vf{xVMdz{tZ*axki+)QiqtpD8 zOT;W**2o!zBNj>Ytv@R>#j-fw=kgjsOOa}FgGh&vwx z#0sfP@4FUm2&Ox1K+6jT#=rWm=%@{M))Dzr$f+ci;LNr#Z0SNDu}{F1Nd7axZBxPt zdfRW`bVeM=MSm_YK@MTMPm>(SMvn1%9ry>Fl@MIzL(KBYu=0r`^%AR%fn|?RVSZ>|>x=*&!P<>aLSUBokZtE~*by4cL9u2FS9b|F5uUyq=Bscv5zQ?PyZ zYYczg&iux{-@jLnQ5psxNfxDrR%)p#nu5rsRoWqHvn3HcADgvSxaP-AF-Z-7_Cdi^ zyORd{>byJu!$784!9aSe7rN0rSPk|&b5R@y-4@6tF+#eQr-#$2pW-oj@%*p{#)q=$ zStDvCMt)h-AAy#!g9}N8Ll8f{qUCqAu-?(gh;5^gR9%uNs6#G?=`2&wWx;z)+yVC8 z(1YcE?q102^$$%nM-rgc!$o%znjadN_=2sMv@;qEVDzU$f1DRIrtS?nc!jL>=1Tp08O^*pBD$qiH<3Hck5D4 zNVpmOEI1r2WFUNCokcRe!|`$i0A2h3`pucWo3wW)(|+-g>3(b}u<6>T44a@Da7V!6 zIqRX@Xirq&PRZ@HK=GIcrnd4f;(tM#{{JTSWB z@A|+tx`URv(<2M|tMHWi?s~{r`rI)Kq=HvJ~TQDP7_Wn+EjvI zo3>zAYd_gg`^3vi)Y6ET)Ffbb4s5mY%QISX!3hMLz$>b0_~_S~mL}2stzr-r9tcRs zyXn8f`CCO*AbvW#JV6{3Gc-6{kY zUd_YXJL7baqHq|h!UVpDm8|CFeB^XP_VQui)OWQE*)o7r=cjT&$INzxnm+1aV)cX5`aav>B&J{7 zh1}lrvuPPbk(`|y)k59tCmo1{UlUX~>a49Q7zPk?@MQyZheGszkB?G&LDfa!1o#iK8aTd9^q05Xi8X}pr9VW$HDD3K884XNyzE;| zZ*P68obeL9c<5;S-r8@D6L@-AEZ_M-s5*5VyX*dgGU}AM!c8J>`E|XRHW&`jepCT5 z^zkKw25MJJz~(VmV0P?x1V`TqBY9LcZNq`4d!Jl~)fH|%0kE-PtLM+deGVW>PQT|v zLCSp`d6D$|g_rS(xmV%J>_%wpsC@9vx^C1>a!ms?an4bMB~P08ea%KveB<^GtrX&S z`2zhdZq8sQ5i)Lpule?9({QcvX)yd0eraE?(9x)6;;N==IeG*jH=^Lqp_zex`a~34 zgAV9P7j$>qm*M;In``SQGpO_O`(O&H6-Atty{b9!=zkAyka0u$eIK*b&KMz`4gYH2 z&ebQnw?1%}&)4PqWObkDWohr;RpNCo4UHjZJhW#n*plDe*k9lA?F%ZWe$N>*7KoUtZz~1m4@zuID?+HQQytiE1rH3$23``+oYN7(0AJ-oI9ECD`XM!|Rv z*#237rR~i7bs;>jnmhU(k8YYyH)aSY)>s-~{udK04n{cH7-$WbybL=QB|fKkJ{O01 zbPE^v@}eHpx?EnjcePhu1tym3sG!Q{X7o76k`?QkZ$?4h_mRJb0N`6{Yirusi|_#a zVnxcMpqIZL4SxBM-}>U${Yb7~**l%}s{lzLd$Rpv=^;5!1qSpY#GKhSL@WEF*&ncY zTY&06DIMM%in{Q5Ks{s**f72@st!+$88hF}4{U`GbFw+3FCA}FkfBk&A4pN57|ds6 z|HxCl`Q%5X36BY;G~9dEkwC_nQT+bkvkd)x3@%Z54b{5p(I!ppEn+BvDOTXVI0s1X zoBC-*6DZ(W1GJC#wAfq_IFo?6A6JO&qExi>uz@r)X9ME0FZiU2pzKr=16&QdpuBEh z^yiVAmlNY%+QP4UtA2!=*1cXykbaLzMD{v)Aw`k+xL`jKjjkrW{*f@A_|aLAxvZUjw>M$oP9H zk|r(gVMS*_O^?AZpSR+D$-H%_(egtp@_e{d^a_6U^Pj(SI&-AV|ui=nXD^9ofCm1`0z)-l=Z^{{{DL z$b}Aj@+=HK@Wtcr_GwRXKb?)NTAoopR6c)006B+2wI;___iTRKGe?&&mdAo|YcbQ6 zlTqkuwknw0#S*Ir5C?AhsVy#mz6lhBr9gnnKsHG{8fF6|`hoJ(H8dQR{asyn#Obcc z0&8&OA;6*Hm8a-(LUH*G__F+CZ)9hjV%UGO=)%tX{F4HmUsbgzg$aCN4XHfB604ny z7{Tzj7QEit52D-oYAGn^fKf1tj=eGnWLfCH`kfV#;0>!8IK2A~aDIp1gk7_|31ke! z6kK8)f!edfp2kf}aJ*i-%#O335bSoev^7neW~<*UI=J;(f*bp>2P`|>V&8*(DG_d@ z+J?CUSz`VgL@5PotrF{7H?B$m4*dx-3%_9OxV_ykv86}-l4CtDv>!`Pq@^E^R_1LC zEtPcr4nAvuOO(U_D+Qk|ng%~0pc-4Nz*$<(;8iUJtJM>YAE#X(|$ z3E83KER|8HV*^#QTE0UKa+(nlqoMer ziR^6Br|$N-R+HriR+?-gh%6I|jyQwEjCyD0Cu{j4=)EcU^Byv_-gj;emRGo>@Q?Bjc0;Nu{%*w(9aE|2MqLbr=O`l3t z>C7ET3i%8%>Ew7BR+LMe9*yOqqiLYI0I)WTtR7XMyuy&NtUs z9lVhewYCVAedYvg)oF{|ql7}j$98zWulu?ny@}gee&xzlg7v2ax+H(D*TbtfA#TWH z>mq+3m(8=iDyOtZ2LHGy=1TJp#3KsD;0w^Cq@DS@$DUc&cqYulWGMs!^4Ov$v|j)Iqqo83oge58e+d^n}O z=U%;Ljlo!0Sxq1A=H5le`h2tL z@yZu3VgWM8=c|j~7kKdLqy@NhoqFi^YIeKIoiBw!KZ?3oBB&6uP(0iZ(&k*!2h-9` zWK~AZX1=lJN&GL`m%bk;!FLo~0mOGmj3Zxf0HBkyQ#+)ZQfQ-w0s1@kG)|XMJ|N$? zxzR9kK?t1)ZGafGa|wwHAM2zm`K{d+#%p=VmpG6T#QTi!=*U%*(DeBu}* zF$n`vXsea>t%(z4@T!4eFLH{27w9t&Mshhd{Rke52F-FxT=#17g4_V`me(zVmO~TE zjX(%YrR$%ZrA^`4O}}bC?}C$j)Lr%F-~;Cf_N9snabaIVSQWBI-Q**o@E!lX?FxwY z3u+TBW7fY7xp!3fhAc4RleWE%TTsXA1mHnK+FsRwkt<%591SP=QjC!7rp^i@3=?{X z`{85Z!`p;Vk#aV!4g%u~4~jmKs$$}L0rV^^L(L%c_PGNG+ zBk)we0jQ%tUn%_XGoQUE9yPH|;05BdT-Hr8suo>F{D0JBf5kYlh^PGzP@VdasVzz3 zlcp^SV!28w^x8xAq5t42SN}>d@>Lytt{5%$@qJ3-hb z!7?TRC-Lcr!-zqOiNTk9N@7Bf-j<|13_(wC>EVeqm~FMCTjtf%Lw>K!nB$*sGIa_! z8aGC0f(*poMWtpHrHIg^4qV|F{}5Vm+6=b1keB=JJ#KI53sw;a$ANbM0nm`7s0Nbh zhmsp~`NJp4Rqu!Fr2~?o4C!ko`x#$LAIiQ!KHEur-JweK_P*wAlyV8qg>6Dqd$>Wb?)QyRihI)q!3psmc`VKH@8XQEKE{skcL7(W;j`NaG6>3)N5{L)oOfe9we!TLa>ZN^%gPC7@FFswdf^m}uoRn+^?$Gcu zE*;R`U%vO-cNDf>`JJpFXLYzW+c>4%D8!3px=X4@o>e==Ld~ge7MwvJG@+c&7nJRa zTHX|RA|i8yW;`g?Cv^e5h|d>Ho?Cr&<-EqBCKD0ElU@(UVdK?YHYD~O@mvBx<{}?h z8M!8S2|e@}@z7g-b=gt{I^TDLK#yxTAHBnCqBb5S##*reT9NwS_SHORWj_c?+ZJk~ zQ@Fk}jaWgh`;1&0#<;2V;(}PS1hFBZI11zAWWvl8Glp^$p=Cw@y&&EcLz88>vP~9^P!8`L8`GBPchE>3H;|>(#8$0u5d?P zbI11cmGgxL=?LF%H>!$Ub%N)h#w=T*3IwuRq8ea*O_#SCHl2SbhRbB$yE5)+7%kih zeq$&FPENV>=3r{zhws^_ekB>#XghK3T{o#v`x@Pip0vxPoNE#Fy$vz_bagk2aYRZ;F$)u2$irbLQWYg$fR-QJ#HCAa}yNqTn1PxxSg z;3co6SwddFg!C>*WOdt>Z@?EFI|zx8lckA>Tn|O5btF;4s7<;PX{+|1)DVjc0I|9L zK-wXhuEX~Z_Z&*qqyBZ$<^}L#qX%yJ5g}e(9=I!L>JamX-&dhtH%yfz#2(IlVJEgb|6`4Ky1iP$%gK7bhbTO-al!i=AbEh z^@#0eUb*AR!fnWPkomc$t3`Y7vMX*`sp{|(ms5UV_fKIqr`?O00I25kF*&l`WNru6 zoW{;dD&{P(rL`L6_piI8W;m1KLETLhaeTO1~5;KIkR08^A(u(^=&iOo2k1v6P1WgNL7fw`t_6^PHh8EF?jVH zbJg?Lr7(1LzjE>9)(dIDn4$be%eu06G~$9q+iO8m1E_qA5Y;i!TZ&7{SYqG=YG@uG)ztOQ$P@O*OOgOzXWTVaQj{lSjx&%Q?JEzq9kZK z!1lmT0QMo^is%u*#6>k_!}hNixwbO=-Qd3h-mEATZ7mdHEe(nPnnRA=h&J%uS<9+Y z#3pjf>533o3vs-!6Aekj^&Lb5@@%(NsUuqtT7^Fk>UI^h8Q~Coiyn_yT%To;2h}~D zxU2S$M+tr%*KbiQbrfL0;WEz7JNeAfq*U?*Y}fzN4Oe=2zJSY%=TJV&4G+91-+ zK*1SY+TMnm4pdAe(VT@N6ye+&ti(__6d5}}`=yuD$xh||UDSuGq(g`FxT@dKmzFbv zldCkB8%)Ng1NvfexzIOIzK}Z01LRTPm}ADyh>wnl_uV@=^|jozCkEw(iUe=KuHj^b zOljXG;TtVR8aZ@gXf!=IA7$t)P(E@J#sM6 zw|A~BcRMQY?4;bKP&}}qsv<3KOkZDF3oL};{?L3l{4%2i0lrCt%h_({$056%B97sH zE3cmBWkW1#u9)!touc_TH;i|vZYa}nH`nFb7l_9ls1E12>=}+Q@-&M`SgKGOL|Q!CmkS zSC?EO98(YhSpjz?9?+pUfx!|e^?k<7u6x#2k%p3(IosQUr7Jvh`w^Yx@=1N49U7Tp zf`$mJ@mR|WICy4Y3*oJ$1DMENNG3a%95}URC?5BB`dHjBhy&wMt z=3u&#QV=-elw1^R5`6pF=a(O?6r&>9>97B>_h!qgV(Yf>)9i@GbUI)Es@B>@d2VbxT)R+0t0zt}e~Kg7 zJ0SL^EFSp`y> zXQeedV49>^Cx^OnHIliZYjCIF%(mGsR}A)|9vsKi>1nHEJLaol*B9u2oyY&ImyU)uoEnKfyi}07TmwsTLKsg?FiW?E_j1S>W>o`=Fl4qVL|G|6n2Vks0a47 z6h>W`;0w9V9zyF~L6sU==TpExO!L^bTtKeU2j|PfRYqsp0U=$DzOG}A{LSwEG;d1) zq8884*liZ%dW=fZ0d;25y$(m_fyt^1vm0DcP=Xoj9Sd2}pmS_67Tm2o(*V3L-z2y7 z+W_Url0UZlP1u3G(l>~9;nmgA0F%vOpplg=kP5m;Q-NlRW2_-Vh$0hACFC%3bvf+z z^q%I}^PUNSGKl@w+FL2ah6QA-)U0k=2p~LfaSWC{qJawpf)*Dsf;s7tdX1jWqVTFm z00zh&I^!4$bQ!#eeWExEXEA3~XQ`(byXx2DCcl6WTRmOYs1>m{eEXI_N7ae5m)*f? zP{37m_fl)XE!0ovI>zVM9+F&w9D^C74T|T@-7w8OVR7vsN1A9MvMn3my3s9DO zMv8QdyiFCLQ<-0?KDu!EX;TY6Y>SkEifU-@zIt4k=&S6|cmst(&_G*j!D}<4i}ttW zqDtcXfah?t2M5tL56{Pa(jG-yo$l%7#8ZToT7yzVfpcF2q!95!aF@2|42E|pm}m6I zJ~s8-ucls-qtp#kdBD*D?Zn$ru@@D91=ct`8lRZ`20ALxceC{w?A@%LPPRn9zX>B* zA1_*R;hAe`O%AA2@$0{&*CzQ!VT8;M;=VGcd!Hx@02}PhPg5M$NnI7Q>Iw5kzN4`*n3anDwa#a z3a)c92RZezTyqkco%|H{1*__jF7JEqOyJT0XRtFOS>G+EaW0LRN{PQXpLvTo9Tuc` zjz*VQQ|Ko;!DZf@qwdyNG}73Dx8CUBYA7~P6gcMISps086xsCxdG8H)5*{$&6EVa| z{bB@&^1TXU(0AKp2m0gGRngiEDJ&TZL7;?TNYJoqal94fGc->@u7q-v>&9-Vya=PT z*Jp;+5>K}EykefFF*xJ%lq){Iw7r zZ-vdG6h}xpyZNxa*_OgejuDR8Tr6*;jqRKcOrq-P6EN?P=v7yMl2Vc{(c!al=E(y^ z#cg|RwrFBfgx)z_$pOi7L2gIOK!-XV)`A02I8}36ple48o`fJ@?~yF;yr`(MS0f2I zxacxuc%V;A&r@5pU#LMNFhfRkETOB><>j1m21PJGq7H>-YVNL{=I0B#1iJ)e(!@@s z&?3&AR_c6FmJY63eqkvG##qaPda+K{L)DqJ=3%k*5ojN|Q$-p4#*{!lZAqJr#LC-x zAq{n4Qb9L@3i1SQU*OZCM-1tXFdDY=DR3s+t}YV@7j6shd0#~q0YE>mSYT9m-X4bE zkD-VgF<+){+*bKIe~@T_=AGUbw@TBmAyaw=#hM6mz6Gs^-h#$51I~=C40Y0nee4*Y z&3l=@UZj0O585C(p+y{D0}fg0u5iDB6q%=~*Ca{9X_S?kEtW#KvXV1_xV||P z$MVo@jEonID1?*_O0A`wT_vQljTP8&3PD)v+unB&j-HcZ2iBO}uL#!~_j*N9FIHTG zL-cMvK(QTYN5Eq2tv7RZJJ)roJfXa@hbQZ!!0O@|hh zoRFZ23dSW+yMycIbp}%Ww4lw?wW#;eY!yw1D1NK(#+LvfycnbAi-!1BN zNUFP<^S~16!wCA@mSu18_Eiys#${kv7CFWgwBUAypwdv>?F=fU;ohf6`XcFj7IZ8? z^y92T;K7~=aRV-07K-cu+&_BS<<>lkCt5qwZYS&GW_q#eIgI68yY;)OmC`vzfTpjR zCJN&Tl?$JH4#gU1$M4hYW`oHh34mGOK#mw#l9YyVq25Q-J%oK!^;EIe<+1b5@bDp9 zo9_XJ@E*xgv!dP`K)VTni83sX-R-moFe*UjH!5cG(NA^)TBoH$#>O}ly3moQTXZJ6 zQN783v{2@_EC>~u>_f{aw-a?wpemB0CABt)i+F+?IZ5aPLt{3RXkwuRc_L>HFJIir z9>ti2{)*W-oOrk}dx$rQ^buauWqQrmOIZv0V(0fZFJkm|0f~30lY)FfXjdFPT+D@) ztwMQu`vtQE!Ce)*(y^Juqtl`r12wgFYx3xan-e~%^t^9w@eWX7Awla(da&k#wqeLM zp>66FP&#WcL-3xX*XJs@ZE`MBxAzM_f_l+mntd%7MEWYU?IPU+#402&0*-M6@%e6^ zuFt%^vBr%R4nwx}=G-1K!v{2Ep#exJW6+%*h!^0Tfo+VV0z4bJRnTyPo&p|s{)H9_ z$%mgJ-NzRizxWEeJ9tFU;>T0f;`IP4T?Hs*YrJ;pEn{Icz_f1?-mHAwOwd)ICpA*B z-l9IBH3SdIq66qM|Ar8Y5jEy5bvhxZP)tB)#3u6eo z{ad&GZVjeS2oW8bY7JVhVRyXb-6buuBBw5akA@_BBOLev9nTDz$W#C(PaNmGp}?3u zrQ6U+qB|~Z_50VZ@bG5Oe zXeFhrGO_u5yJzCE(J^nVJncQ7@=LKEcHEd zC)(-|8QpoYSrli4c&{4NkLzdsyx-GiCs|+v?9D_QC&QL?YfAufx$(~7B;9qi8IisT zy6Dc>#v)RAg(CS$t9sFN$tQnFzjR`zP%`k~g*>xfjBW!Tmm0z1O-@D?4HlhE(80UU z72COyP1rh84({=mST$SZ!ctk!Rn10eU}^327x79m)v48!)?^~mTDKz6G)SweC(nk2 z-%e-@eO_V9y1x^XjX4gy>>#C`eSw?W%TD4Q*B~abBmsQywtbp7dofIqhfzHA!|c`+ zv(f2Qyxk0LqzGJ&s;{3{V}n3wrKhgmM_d^8a5T)d3z44jb%j!8b2uR2gka!#-*hMO zHeZbTbXJ|a*nwUbDi7fh3hdK|6i81a1P-3nx%DVWS2;W=f^N?e-to$wZ*t6C&c(Dw zkDC(Sx%liu&45y&Z>YinYR=N-cJO44OVPc#gh@R&n;jT(r(oS&U)r-Xue>z9fm+;S ziI4c{*$1QVxQDzEYP~N5b#|`9yz(zQD{qntqzHtHl%RARG{KWh0p0McM_A6(|KW1c z-kl`Oy;~}x6%aIVQ*lG@i-hELqIfwE#(p)c=mY?_E9tTfa_+R-iifB|CPgJS$>XL}S8Z+}LB3d81L=3}v~>r5f}k%95(k7~ z+^+;aQ?@MZ9FeD%MyrAqHL{owIg*=Nr#kdpltQ9z|}S zEmnpFu_Mt4r!GsR_pJC0v-E%}B9ihCm8)tCK%0@T_e%lXr_e2E8jzVOLzp}`x@(E} zEheDW?yVn!sxxorjfVQ)KEh?)g*!>rqh{7g;({-u04ys69HQ26-5`)R&*dfz5F*K1 zfqtO_#8)=g4C+3^c1L$4cpcU^ZxZJrds_LtOv419&0N^t)wSDLzzADF?o{gK835&h ztoce)Z_3^4twy&ppJ^@x2@e2t1j$T# zNe7X+6_|dKcG3@clq9M_m$2pbV8b9dPSwDN11kOvPv{#Ps5V>{F82J~r)gxSbJ`G} z;p~eEwN*@BsLMs=ilU{Ol$w zx=}A5pqyZHRZYWub&`j`gd`2~V3Bs0{QsiyZOCvEbiNInF3kHqpZt0L~(5&MnrALo`o(98Fu zzLKcFe;kxu{J_xfo9}UQyMLS%RPEk32HpFPw;vBUpVvGXzkVt-sy9xXGQbgB}`TIL;^FSN-`x}4jk_lVqL8BHm z#uB&S^L=c)g2w3WXrJ5cq?ntx2rY&+BFw4?@XUR_iQmA@_jZW4Z-!^`z5$9r?F`Q# zlp?s{|MwHa5d65umu>Io`+1~!JbyZM3io^dhI?ON;P5okpe3BK6SNp&%ET=;%(Eg1 zgPi|(qX3QHeC5ImzHV&Q@0avygBsh~>QP4VJg@{_z2DS7Vfs-E<_8j&I$!PA`Nyz7 zkl&eqG^W_R>wnJlTfCppmpc_lx7hlab0F!zSGTGR^36|Me>V0%?X#~c+?uzWo5p=V zT5nSnfTF&_J^FI%<00u!AA8QfpXb~71B0FBSK@3L=DYg#Vt&rwUX*z5=MR1CIKP}f zH%}$ZpmzM}V{flMKiBnbmBrqNZJ@FDADb+F(B5K9-T&66(TE7zCHc8j&JPgeEwf%cy|x#d61JgqxO^G*#{S>yH#D% zC)I43KsDP}1!hb>zaa!+c6&2B>ng8cu*4|729Gy``E%3N&kuqSbafogQ|Fotv#El& z2P+C0ocFEsto}F!F$($s*8Ke?g6KMir!ucK+#CTwD+QA8{?5Gd0)OQ33jg!CbrSUw zM1JOiyoFl`)XC_h->r`~6)8ZOJ%Ke%J2$RF{~kk ziM-8`kjIcQ?AvIp7ZZe(KaYNdV9W7^Bi%I7yp4+ax*t#j{m1^!M}P6WxrX{9uk#{r zDLj9NaMG~B^NDO5FXL?zM0Ihum006VwYZb%8bj=Mn;RUw0Dy+WwN3sgVEmCp{;4yAXrb>74C)~xbf z62f(=2+cmzh+f$vVS<|{Qh+J4UCH%a^6qGZal+WV-p1%-tbTo*ur~r6>?#U;_&!d@9nJF?)nT&{ zkh~M_J7fW(duSe#L$y()jcQgGSpXmOF{hVp~1|;=wV+B>STMg=63N+n7 z%4#6;_hoB?em2r{-8_Q6Xr207xA3WZ9h4dY0tOmMH}+QPoBKIH0EUhb5ESHoa6-N_ zs2Uyo!T|uYMbou9(L)L4u`k$HXKaS>B9DhsC3?dlqsxiX%3$#ToIbvkz_v8M7$vU4 z1Uuk~QA}JTZ@k3GQ?(M(nO-W_gKs{sl6maq?KKB)ojdrQs0VXceHdO9ojMQP zd=X@IRiVR!BJ}QZq!7Qe22}GJ>Z#pq1q?RLt9SdNeDM1)Yv>En=zhPA^KGanAyvjk>14D5v9`}OKYI2u=Ny;6(O%1S;ZS-6K@Lk4%b@(h+=L8WP zDFZqKmARm2BmWudofF1Pq4O$tB(5*+%$k>WrzbYzm(JTBYYwtEpkS4{d0^p*N@7lj zw+(1Cz^mTJ^2@&t*7`UoF3KAB)3xA^(XHYdG2Hd=5ouXF>&*p7;B=0sptlxxqQ;=_ z;vqts+l*0c1K8AhcpO+bd(n}i1fdy{i8*^Ggr%t2OZ(VTfgVE%%<^96L2p-Knr)IU zj;9!cixj~jv_FNdrl3@H%FVKRY5t?~q`uir%&fVEzM`6G({L{rkA_M$Re0{4D40aP z2`jsJEJz4R?`59Id(}tp@|`VNcbJ2y*P0EHY88OpKyiKxF`HK@14wv{8Fagkl9Tcy zsH|L4J3PR8Gnr<3A^ANoJu9^69|JRz^96{wp@)q=M{maf-kSLkVF+uas|Ii)(d`@s zB7<)atV0=sz81KE-DjfX7Mhp$^|3mmigJ>IJ`X6Ux^g1`U$ws=(CKYh9ph@zn%q=W z4*9(TB*kYp_Qqs?HWTVffF`qx7^O~jCu;EeGxZqkZQ~mcz4vw74-%j7(j|HBwTar& zn(ObjoxjC8@nSbFv5Ly;9i*s{qi`CT$*E;dY4p={QVV| z(gn26&`JU3vD^KQ-nQT>oa$LLN007i4*wXC^Ce4BKmfpnHEZv#JQNnCR$w+0ocR7MkCre0pJC0DMYyov^hGNqTtC^au(xeZAK zyy!6AWYUu>)CB`8<*kW_=&G{zz{v=wGk5Ee^YNWISD-@(GJmpegm=ic)DJhVS=rty zJK{~1kqPahm$6ze?c`V- z6WaaYUg+&`#k>~I+=KIc&H__|lSeN$Br}2tc5UTTHH+j(*v{Fn`H=-t1H^L!4lqKc2WikT9;1x zP<#5Cu;$IZ&rRb3`gkG2*KBh5R43sQJ?My~kjelZ0@H(a<4%G~EYYa!yj;WNsLNt{ zw(+XRq#QQ)QLCLvvDRp;2!lG%=a>4VwK`@_+-vDC+X*dpm-Y;W zN=ccf2c9PA1LNRwIC<<6Hi|1lo}7A!sQNqV@POZR_l(l)crY~JoaLByj2xs*OP|*( zbo*W@^s9nXuV;O~nQHSK$3${H*f=nYc&`qc7e_szATkX?@8Y z;SQer%&)fNkq_7=%+IfUI|2Ub#~H@fVU zga@P=FXN#}2q1h#bN>F*90+Z;pl5rA?bZmnrZs?jW3<=?_q}68_hupwM!u@sA_r4% zfJmtkYwSP*`$_0;PaA4`x9w~lmwPRYI5ki7m<)F@6XLCtG5igIm5%;^c zl+OC|%6d(H?z(ybSm<+ZF=?y}U7^>*f_-?j4Yo|^);5lLQ@S&1e)g^Gt;!zDG7O`Y z-=w0@xa@FqPiyhsVUmzOxUTZUAusb`M4TXjyew~ztJGsq$VRq0fEnFoP}7czb@c3A zw4^S~`BCecV-Ns$_PS`;D zpbVoK77L7>(5|O`>Nh&{Rqr0X?olOr1TWtPCj<{dFvF*3bA46vJ_AATZfoAINl+GN z&m6ho>}AaYusEYW?b*JTbCpPxBQBhEuHs=7l`@U)dUf`_xbgbvWH^8);ZUx`OxsmlzxP=f%T|1z^mV&Zk>z=KVfIAK(1wuP)$B znLCrExV4}Dq%u3i65@7IlPc6Y;}KE7;W;h~00Z>H>Xg@S7;ggX`ZvaC0>*#zQMt(a zk;}SDrnj%_(57H}irX8+Sofo$PkOpEHX3lY6(k!nWxHMh4uw&bpfCtsjwj>Od`U>A z0=69vn?+Tx9cFOrOFBY(LIFz?csc$^m>bcH$F5C~PJ%JP~ z*af%j8XNKR@J!0{jsrIPcpn~_oR262uyZK{^NyfN_f@M)5)n39%L)ek9e5|;dI1zV zdTE14=0ebp_hh_{-SJNkio#1i?!#?ghr0Gk&+VJKk8s<@L|20Wsv9o)0u;=KzP9MY zV!ZgIqgN9?vAbq=HAkJ#q*+xgUsKuu$0AZ{Kog!Erf5pr2)mps_ zBoQeS^I`xfILK$0B$vvhv#Zxuu&P9P|47uyRP?pou-aZuOVW%A)1oEcl#lmP}39U{ZmR;h)yU^&^u*BV&WTi8Exvw}EHRgQo4=f3BhKG0}>;x^Vr^A?s^pG5H`EZn%R4hrCz!1-`tG_glO}ShmqQEM*?UOT+1Kg{L z;@CV!1X--$v+A?3$2I%F@i2VcaIIM~E}wG7mOL(RS0{1Z?CN0UJpml`v75QnAWB?k z?TBu?a(*ujtWMg1H)&wvj3(D7z>GPzH?^GakIfyDoZOW!1VboBS87+Ze3Mk%7+XvfX9x# z?CPRq@6rQ)Z+N}V1#R;Vw;#Nh?D+yJd|xG2yDd){^F>Z^9x^q4siz~9>&%sMu() z_^={bH6P{UL429TGa6B9fu9-85jI9&+LgHVQo)Io8}*mN(+JN?5{VYe9Gt`4O7Vn- zQ7Ar7?ZbosBIv3u`?QzgXi^tUKT9xq*|Cac|oGK+Al*aP@0aA7 zz`#Z|K40W5xP0x*+WGz=X*N?x?shvs5zUSyoCSLHz@Cz7&e@ofdwNVHU|1^3oTf&l z>ixw!4_-145$IrE3qM?A8c4mabEck0CFJG?z)|siE7B3?zJ73YU5z+vwA8gtun!#$ zBnAw2dr4N!v2@JvK!bJ{lX}ADEY}yVU9++QVBG9l`k-|_h>*!4Cg-a4lGBe*0k#h8 z`xV2q0Rj5Y!)tm4o|Lgsq(za)7+SSjopzkqG5aot{cc}pL-Y4EP6x?{{p1k0d*GQb z@Fr%()6Ws}kpK{qCwH;bsW>nn0wF^k4N|z8pFAR(y65|DFpS`_Ll0XGTM?$nSl~ft zNnmfHt>7l&(uH1|9za%7>?X3{9>`|h6i_la{9?JS?zrLt{0dWWLi#ngl$BPZiD=h2 zmJ%_=oY`xzs_vE`kBv_^>%Cvp_kv>^u<&dzbZa^TW$k?o zUm74R#Q%Y>_RerJbz*kdk2&BY4Y^62<)*$MKs(vH%gaCUJe!eLszc!3uAO*)iQHYx zL8ko>+G;7IZL5|R;oZ4S(*~zCuAtE#T;Xw3M=^ca8F^+S-9Y(DK%gB3q~pYr4pH5e zla8c+SeF8IUb&}q9Qf9}-(OFA*Tzv*l4{FJNCViSV!9MaX112-f>YC}`UB;(4ZAI@ zq5$4rSJ9s8ng_<++Odz}O=q$&0YskJw3;Hdpf@uq&b|sc>sg~kMVq`7Auu0|O1@(n zAR1`ttezJuLt~>m7fg3MtZUO;ct$gXURt%23(K|{?NKl(8FLjf<_dZ8W_phytJlbP z&$goUP9fNu>4U$k*GrjPx<2u=m!%sWZ{0=T)jm_Kt4fB|s|o$&vfavjcD5&_QIJQw zi0!ylZ zp)d+3S`JG>t^}`MB+qL{-A~zVN5`HkFIHEw)qBS|cMoXm6xpL*yap# zlTAIhck00%Qk^#ss??<429vSh?gLNQsMmqt;66(pdNlGcsHO`v3#}m}bSJ(?^3J*E5As z>oJUOG^-UUcfvZwGFR7zu@#WrgS3?u&bD1dO^xr?;h6(fkYVg2Xb0T96Un;WoGb>K z>$Cwiiqg+JEBR$};nwlcc3Fm&QY4_jLu%c1h>n8sdCgRs0SV1fS*zo6)U~ePUSEx8 z0?a;n*Byhl)wZmU9VZgqE~!>|3m^YwQM{TbeM(+dO?JqfBdw3^GFvpQ)`EZy-}^$G z!El=`G;-^aNYZrBAFn{NUgk`qZUvOE(I!H)wz4vNlJ$;zy_ zUX>-Hz`rPZLYt45gb4VO@mLb+GpRS-kROeSpd;HM zi8Gzj{?<|%)wu>5vNiP1FY}`H(&O}~zezrA zAC?2z$q_B-olzWiuTwAk6$2K(!~Wd51;H#3pCEjiYiIRX-vw_d9zd9nyB(zV)97BZM=;u{L$`EYY1PH##8WthQ zGZHB4cq9s-y)IrK5G>VK{xw6%u zW_Hi*;9;JQ&4xj_kYcO{1&WEM2k=@~#2YY4Flf%xs(f&{O=@1eNu{?T|>)cgb^S$W}K;nOQd6W}}gw zJkz`APYIk9roV%E0^B)I)rpjmdZXg%d6_QOH8%tf?t;E_y9-nnuy(pB5b3mI&)4X| zs0fm2_OS=3=Nc1RqT%ao$ZZS+b-R=|j^nt?JNVVXI6sHa6DV-kS*B9M@yEe%gSGcI)%& zCNJD5B~*x{MlwiNpxL`zlK>@`kn|9t^|UwQNNub>&cg!)DbSf)v&T<59_z1a3@Pl8_`%QYA6brn3rhfZq_kvShm<|HW383ft$y$C^BQ4xvkTCp7>kS=+UosL?Ovy&u?K+z z-XP5J8ni-jkDwimub1Le4SxM*02@NFGDh$UuuAB@NNC$Zu+kwpWj@LLb;?Y2g|aN8 zE?453$#lxt86d-DQ#c^$i>%~{Bfwf0So#ZT|D`Xj_NOy*ILaiV3>;#6Ot(>j%+S8I z`+8#p3bTA6i_Dd*&5PSsf6+z?zWUk2>xi?$Z)py!YW*hOaORyFYAKgsad`mw;yXgB zq2z|uUDhY&Ikn(OqaQ?I=JLZV0@`MvLxv3bgan$t>Tg#>2Agg6amN?yu&H3NPj~l%UOC9g)KIIMGx3)fW4L8bZAFsEZC#1lAQW zCYCpzgWm?60|3x~p?T6yxS>?mQk4Xq!4QCTKRG4HD*L>h?%q5g?EZA$z_!=zd{pERtw#IZGI|noH@i?mJIi#H;dvr2vh{?MLRvceSXU+#B=EkRp~Br_vBv2O z0~O$+B(U(JX@9P2mSEsTl~JQO@!DT3?Pwh2-BO;dloYHrsaMi913wDz8B!ECr=2|q zHkPR?-!{=~hJ`;HBK*8U0Yq^;yY zj`EDy;-Nv}`L!CLkZ$B~&ZP@TZq{qmVl~z2Q63q$P6VGo{?0sNH8R#~U* z1Ws>v3e;Q*w$UY)$W(1O01ow|3ertyYZNsth+M~BQQQ3SEN z4+axbJU)Ez*Zc8tcui`dk8`~)wWaGwtL^GxEIeK}s(`MSF|b3UQpt7%m(!q$J@v5G z&<5d-*Y*gvuG@!~?(!zF_);`0VPDr=+^22cZQT1X=C83u3*eC#Q_b2kyAZH%e3iZi z&Ta2MaGDXi#8ycrl1ozWUnqqm%ftd(>fUx}PFZ?;`C-pjw64v_+JlPia*Sy!fg~|D zQb)88zx~1e$CUZELTo-7#up5}c>eB>c5#y9J{!otKpwasyUM=;ZhqSO*;xMPBW@67 z%~$iKA^hJ6c>sEWT%hs$C)fi)EW)N z6B_XgFvgPj_Ssr;KP2oB4b`Pv9Ml@)ty9!1mAA@q17iJpf+uRL!6{q5d*k zf873M3V#n}9*zsk4&eLjs`Na&gADg}!Ve)go&uR4M+h8omznBm2;KC?*exeUQ zv4xlR=UcEcezV_vtS|Qb6X5gXl)+8=->HO(D~9P~QUCc=B3B=`9CB zh|36awbctNR|v6as71d$^g-MIa`PAS@h!+S1PjX>;&iqo@89`*5vTw7a`@%8FyDE^ zKezpRAcP$K?K-(!{6KwJrcV>|&tEEk{-qT2%WwU=*T>HV8q*<|pOVkZsAueCH{a}E z$1%Qt{_DTrL0)pRT-ZD_KOJf2V|j-;`vb>+!%V)P=sd_zQIFw@y@9l{0q#XC4=a9|@EFciYd;$@5QOjQC?J6%Q7TpFjDYEq=@Z{q^qJ;?H@~ z<^|)=x@#Yc9QDI$L6J*>ThIicpadEJ4GF5R%O3lYF#GG>)sN---~B8K)SqWRfA?Me zPQ3RQSrtE~+y+7<>GX|NMLQ7j`Ir%tm~r_4_wGX@A(E{3qVE-|c$*WgSZD zPs*D7VTbb1xT}9zhjNZ}ek;xLQzc+d{{M!(je_p}W1-+@39#!bj3KW5S#tC~cnRm0 zfN%3>C{JEi%mRu1-=8e|xQ>5=^t7TE_>Id)>|Mc?_$b3w9 zlHq=2a{Zjf|NPn(O_D!<&PZnY!z}+k5BcX8|J~)SKf4A*{f_+FPmk5_FtIKeslB5;Wj_nMHBH zUfz!Mr%f4(T>bE^#X?f)lY08gjQNZ71yP@#>({vH&!R)OEHL^In=-5-5S0H!e*PyC zowrpFis(0XVrn@XygzT^5g?6z|Mt(NazTA>=6yC%ezjHqc|7RU_wtwDOuRozssAOZ z_Q|t9Pn=&T?*E|O|KEv+2k*arrugs?ldL{hm5*u1epGqB7`wm7l>1&b_J60`TjV^O zw*MSk_wy-7(I5WoUvJKRFPr?$ocpO1@cXUSKPopeFRtGRkf0&%+eP;M6yJJ&{{g%D zxrVgbAqSO zyxTXy(>IUmzr`c@W7F53W2D(Z4Ss9he$&7IpMF{V<5nQ@``7*yzVDzhpSj=< z2aTb=C(-r2#}FOeKi8r8^OG2#YyD3PBlT~@3(mQ)pRj_zE|K(}&hWlS9sPtIM5D=< zH1pe0eF>SkXwj&Ux353B3BKfPAhEe|H-8{n_~rGGK{1 z%eRZyZvX-R?iBQ&oeBO_%lc+jo5%XgO)|6aaVdi%lQzL(ki zW7GYgnDAw0fru|N4SksJufGHq%a6(UHGcwz-1q%Hr{sTtnf%X4p!_j+#p;7H3fky5 z7WTiQukw~A?0)0I{x^7dgR(QLk3!pr!KWzVukEsYWR||z)PJMP@@*yj(}DTTO87Sp z%53=GoR^wS&L((6M?SH~!@ixKdo{%3~Jbqt;2=N2Qt$y?2_)a&-zkI5H*qp!2Oy`)N zx!nKMx&Gg}`YW3Jf%x#=K9?Ja2y^Va=lb^V&z!rzJKp%So&C@Lum5%SXTqExo&CR_ zbHDgr@X7fV`-1mR`Na%y6k!Ww1|Bqr2x`UyO`1;HL4;zh5 ze_+?Q;_x*nk{o7XJ~n*!pr~=2583zcOSpft%lnu9-ZJw?R`VB;!u#d_q9XH66!f!@ z{r{CR1M`^+`Y_!K4uQfK%l(P;|MH0cWfGl>k3_?-S?e!;_fJGBzu6o683UZ(`ec#f zea~tBmhUs~HhcZ=ANw&r_~~Qcgap5IaDF@H`D5Pxd(sWpfRC}=^H9?2i68#Y4(+2) zF=F|@eYQFNW?1l%>-)!jr7!dBlLF>zp6GMizb9`B`0x)j@lz-5&+2j1;v+@$nY8-5 zlbHXlbG3ky{pu@!+Rx&8yuJ7c0zOaT?{=;}PvT!n8MF9gPx-JnKhiWIbMpr%V?NLI z7yNTSKmL#R=#L%$&C&epP|_?urp2#Wy3fAUqzG;^k;$C(O~N4G2bVZEs_w1I30rWf z`o1c#Lq5MD1YtIVkIU$@OEikF!6Wa#9D&c1TZvEh)x5a>f=1=-3;C`Fxy@`YvCGFP zh*8i7a1GvHg3h##Aszt7L*izB^96M7t<2wKXaHN|inzvCgU-tvVpZ~GH^U+^CZ?1Xq#_P~g9*BUh`bLxDJ&(KBshI=*ae za@8SC*Uh7|=Ryl*>lQwBhXbw=(aEP{IcIN`zPXGE0KlTOt>Ws}0UgYsmszh%<@#Au$v@%#c3<%*~N>(qMUyKr0VS*j-#3&}N zkvAZB>Zw|J7ECXd>%ljlSIIo~@^+SMY~8`{LCv9xQ= z^5Gghqe*R8%U}rZ?MZ(vH*R}wmBVOs90Uq9-P>t|a|=aWLUstDkpX#7>t(hNA!AO2 zP=`zTJmp6Q$sunbI$MlX&xtzsdQZE>OYEFYbl{*;HI9$k9ny#4Rne*Qs?8?KgFzKK zJSal%E=THau33Y*t{Up8-E4(|Xws{9`=WgC`!H+h3(@F)zs#W*d06|D|FDZn^x7ye z4MAjTs(W7LpzHJDbSxhCg3xMm2ReG!9lT9#+N5ptTyT}8jMG*7ELP`)Su~`4ISf2M zLPI0+pP}A4VayaduX0D?`r^*e@oOOoZJ+c*=WUNQ*NnOTBBgE~*bZ}7%<1s9F`?Kk zBz-Ku{Oe$?kAnhXBk6v+798{fE3Og4T`z_)Eo*1Jxu`{&p7CKpn&M8>m~fpw#Efjr zSlFwoAzBZQvu|!MI$)7(GQXCKoV^pmQq=6FeQc>fkMq^u@?PgbZ&zWOZIUjIr+BY+ z6u}8)dkR}^mcc#cW?8*7|515T-|Qx)7UXCx7B$nR;a)Bt4V8la`MGnV0J<7YSV6mP zL0+MuF7rg*tG=a2kp5=f0i4_r`(eYGEPXBBZe+i;`}3=mo!Yv_3{#&+$w~P!)=ig$ zG&EE{CeutWB){jSXQgxYF)$-JUnKSNa;KV7^LEVE7cyUXoWk1J#*i*K(5-{x6OwKZ ztV0=szJ>xhyU#?&Ei?~`(8uarQN~FM`aGakfxwLq(t`juu^88e)iJIXt;r3v?dpwL zn4@t9N%)xT&t^hh3Fwz}5wxe0-H95!9T66u~ zw)3}GCtmEvC00>+y+DthM;s#5npxW0!;n9!oK7KlqF)}TkxS85B=s(-9QpezE~N_; zBGY@c%4%A}JCPLF(=do<$vJa+?vzg#dZ9bLbez-aSeyOBgKU<7nq{e^CmM3E z-L|Cq*C|rMjeJk(=9MS2gdy3g0_C`N!zGffU$CXy6-=7DcX}FZnn*`SpywK}pt3cj zP121uiwBU-lI86oKU)&l+{vv$LF6W*p#-8{Uauuru6^$6wON_c%VU(9mk@eL!S84kZq|SZd|jny;XL^ zn<^s{+C?v8wO-oEu{g#xq)=6NYf*dX4?xrLt}tWm9^()?eai$N$@&7OV;^PVLDwO< zw;$XKy&bM^T=U$6^L);_=Yo?*FE%7If(UkP1zigE-eh6>!dc9ZY{ra-I1jwrX_G3} zl*!NX>75|ySv)7gkal9D_5RSEGHMfObMzQLkZq2fjGEfUeaMFj` z)7OMGZ|;3U21|fYuEFNbCWlW2o!#g`M=VtjIe6Ha-ZU!jmpkTh--4{vY4K2 zyy`J2ht2&IJ59=n{RGQf&wcHAvLn*RJH-)INPgJKv(1t@Jv~_Xm3C zS2%MnEJg3}{qvBj$9j7W`k1~w>NQ`RM{~4j>k)OVQjuqWbzY1jEypcJYCP|&M_g)- z66ra_Bf~w^La(KhylCy2^=r#mV&+H(wJfYXlzYo6)My9@)7$Z!GFClICx{j1LvT7| zVZr#|wm*<8cbQgD2Rt2O4N^?L^QbqWg?CzrVJvL(usAHlWfY3(tl%N(xwXjm{vHwv zb=6b*Gv8i_3D{f@fOyF>5@IT&>Yq5e5~a_b>HHYjw<=xEHhwA+GX&x%;l3 zSD|gYUlfU69*r?zy6NU6q8igoZ@)fsednI!D3Oxf=qie`?>%re?7o!;+y z0ap2Ih1nzwBkWwsu8?6tM)7@i!>?uJuUI9F73`)82qT^OW@SK@iYo=_+pS5vY<#nR z48D}al)CXLEKCgK4TJM83hY}N9_-k%2cKN6V@L8vH#OQA)F+>uvG~D>ZgrgKo;JQv zk+-RdfAaadCCr$$@4b~T%>&j<)yEof!u{`fjZSzbv03eXEVkj(o|6I$Qi1a6$An0A zfDLk}%m9esaGl0AI%Tn|=YS#c#Z~9^=t1}N# z6OQeBo(JkJcALKgC!5>+0<$zIW(A7D+z6RYAARTZoZkW)c%Oc8f8N>mS(bD0Sc}3K zFvll|?qW5=Ai?9E2odLK-ySFU2@*c??eya6kOul0sRH`sx1ly)LC(8-zAeBoexn>T z)fW2o%as?$?nN8Lz;WefXE+Zb67s_!nk87xsvEZohpZ!aSP$ZL@ail)f-di~JVbum z5D3%#TCU|`$#CC8^1jg7vI1qPcW?)=1*vw7lAne0*)ENiJF4t_fWfWr-l22y$2)9B z>aKyETDqF)eVcvX z!08o?Brui9DxpYz^nKdy286oCo@k^yyvewzY*VPgA|IMfp=juY&o$H4eZu*ye@f9HOt92^G_k=y*#Em zQp{rd4AG#Uwsl!GYd=i=&HolEZJzR-D}e$D4yi3bSS13jJbCLe3Ao)3iHF6aW(_`l z1VdV%r>#HV;8vYYkqYwIH_p!`$QnotU3z=uTjK*7(A>d57wGMNrOt5x&EB}sD{!Mb zb>^uXk%^3=XTqfuu$t6FwdMi6lQE8}7L`$(Rct@A(sXIjrNPh=V)qLqkazli(XG0T zPsI$MU}O1I;3fg^=KXd&ez?FRvk*;ciQo_EZhYajO$Xba7(Lbq$NLJp7tJEH30KN0zOyZWLA?;QxBLuQ$K{hq*TgvfcvJOZn(K}$J0FvW-5e@{q&%&zIVTdBh00uf>!NavCTvf5Ct*w2 zIHADQ@F)Ax0EQ(I6_U|8kWClSFWEHOO#2q3PR!jDu+=83Pb3i_6DPG>^*oe!fI|u- zP@1X5)pBTnkqWKQxH)u<;yrsX`mu{E_4B4NTKC>_AUnj>A9as_WLkf@O#H?|%>V=g zc;tcVZ;ZwSW5EK$yYsus?OcW!V!hRX#B%|%((uZoh%P5$y1K$R&bd;amKgmm%TJ$R z88_Qw>7MMxdh+_DSFfI7BH$mb?I?RJq`r``z)|KoI37M#Kb2RFP1gDhfCP|VUr1|i zV)p&wxMs@tW#M)p&ZfyAYHH;o{q!B6OH`R4aD#>@t-Zd?=hvw82Sx#TRPM4Pr;hAe z7MKPcXR7pb{jCcE66;{QU0|YTnpG)B^k}Ae!PE9RkA8!#>sV%!7G0Q()-X);3=85Ojm-mNOsrsjGN=;)vXTS?hA$+JZgku&5=XIUl0*7Z(e3g0Ug`jZX z&pV9yHa-|!(4VUalw#ktq9fyMGxGSpKKN;^+By*I>zrTa{?1~cf)mSg1UH7~^Y3+O z0r|FXJVP0gI*O)NOrSKe&VS0W1Wx~cbYp`_o@K$K04*mk-#6A44r7YqIj$!-S&R@f+SYmJompXjaQlEe zd|>dR-;{;{@=+mBhi@?;)^;x3zGEB1Y7T2EcV_@%)OeJ7^jn>{v-Du!*g+Q!iQl84 zYItUG%=Wev>GwygGamkt#luqn#9e8X1t~VzAnjUeGQb)$^Hvt{J!k^Dn=~BQlUdF2 zPS7vHJNUU~_Yye+BQI+d#_5+N44}*AFF!Dx*wP)!&j6uJlA`J@ zb8*@+SD#`K4??6!kv)!?t{vt!h-#3X1#bZ3eZW=2*=|?Te7k+0OjzVj=7kSy|BLbG zFdj@wfkP?YI5{)>Ok?~j%M-50B0q%SpSc$XhcjAxDlbFaoo>s|0_j>dM;uFcm*&ikw#+m4u`^Rq4AuVUjS5XjPcJEwj-f?Xq;dlV+r^JbGs*-s%8&pdPLg=U?v z*D)==XhY{Dq#pC=40W#$V!r+uR{H|0R-m)L=RU4+SpbkS3I2sWv2S)_*ck+He>_nS zze5Kk@t}}uV+pzz4Jj-z@W7ZCoyG4M+}M18?Ynjg{L5=xHI9w%__#7$Rn-IhyctNQ zsn8fs8f+T>ZC~H2eto5mntlZEVU8Lx=&AcY!Y{f-C!g$8oZVp7~ZWuV+Fp%3U{u^Jt?_>tKc`X>q7c6dCiweAfQGil1 zL8fKAaW_NkRvNqf(p)5eIX3}h-938$XnIP3(^50`YfnO{Ve%aKPes}`e;gO7yg1YP zT!8$Z_P1@uItix(;Ddho*r!TD2S|^>f!yL2d_G*JV>}8d)*_0;Sp`fPOLwkcX8c`q zlBimJK5QG9&T(+tK~`iy{5!!!XK?q^<+1=Q_H0o3gO#jTBF-1y@hhNNpC=*QfWD2D ze-&JD7aHT0L;%0G?vpe7HS5*UKi|_1Y>F3>3!k_|enFcMD3%TbaF~gj(-rY9>eXc6L1URZxU;Ze?*RjA%p$ZGO!$fe0}yYq~M@Mo|^7) zbh|MP?$#C%=X!%*=xH9P1^?!BBlEOgA@g&PsI2lo*M&HDPcHcFNeLD=@e=FRYw$M% zyx$PM62Q^U6WsR;$=ZUR^Xn6jje9Q;Z!U=H^SA+hXSHuY7hB}N9@i#D0vNy(pH4uy zQY*V+035FVZ!rje0@mNo?K=0@Sz5eKfn(5t>=iNG-kL@lUkZNB#a+49kqJPp-<#hQV>D&>oL0tbfxLfMk1|-YGbQ z0%KKj*C!r@!{UKh9PWU3BU#ahQ5I!5X~gPqXB8Z%K(WnZ`yhXP=G0P$nC|Ysp8`^E zz^o`>X3p7WFO=#$X>&CVn@7W4{e6&!celCx6Rra69bvf$YBk*zjH{QJ`6{^zKy+M! zQ5k*8Jw_@X9ADQmpLG7pvx`)+9#eioI8fiMoQW!GJ}Oy+fkJU5K3R+Jzvs1;&aPPp zYR8Pce=M6!1htGQ_&8{=Lo2DaY50T$*c$PaT(&OjNhmzD^0+2V6UTY@AS{K9};msfw`AG zEpuKFO?l`n40VO2N`Dn%>WA0Ni`cAwe_T6(jla9ZO>4hPji=n$7>)lTt#e`HA%J1HE$wHiH8! z+~1XvG~ir^Q)n%}&u5^xhT9}jS|TYz64}7f`-&sA`l##;93nWM?KT|(XhZ>7_Dj;6_Tr&dc%Xk&cL_BLo9oS6(12j0=m|pH!@3C`|2{ezD}%(;V4~t&jq>+gu z5~J_o^ETc-?EI!jK2hiAyY3b>bk{dN`#+5y!(e^9P#^|EhU6`9!}R;PfCGAWzqG~L zLLc>3PD=wIX1<@Ck-vXyL07PG#Ev(g-51eg+qdOg_mZmc@rSYUhD;UL?s6g7hXG$^ zW-hl!v1mN?6D$h+3%s$*)u7YN1OGZi&w%*Mz-jt5cnzPkgSdp>xK*?RP&i8?a~-FDKpzcbPPwJ42Ko|y z|BgO|ONm3z@v39?>zSTfoMUNKF(2H=JwI5ccmp>=c!N7O)qdLmGj-nDbLPA5=`L%N z10dx^KyEv@TlvEC7!bJM47RZ0UJsi0JOgVbK+)veqG_pDy>67xJ6$)f36GQ{ntk?H zh5?I10Ccg$a;09tzsu6B9%h4jD+FVZGl)5<(7p+r5<$0-VnHo7qX{p8F1w zw$ac1T<7V~cD)>VZ$1|Qn=DXPPln3}nD_&@3wxSovvT$1W*9zZm*>e5z^hh<*Pa`j z(J*lIWUuk>4ZOpT`)a=2mW3dje0Y8ly0UO`;cK|{EZpxu>74WI(Y6mocZjM&XY8QqWLvD*1~$nS<8Zg(#V#QR>4yqgjP_|1Wm zec+51B4dqPV4J|Xv|A9&oW{jih@0$$Qego7GdwxXxl7@L(=jY%r&UAA{};YO?r{zK5EdAHFgQ-_96moV zlC>N{gXRan(+emffoYz~fJ+gyArT(m)yR$mo&>PKdrj~t5|%-*zOi zbIF&B&<8DY@Lt#t^4Cjxy^^dw_o9U8Ka2Sh!0NLCw|6mDxpMJ}PLdZLT8FF$9yvnP z>yM3mzez4AG~rGO*ZgPYIeZqq^XnX~vu*6Vyo48KIp_Us?wj{4@ALh=e-tXnYfQ6@ zGFRxJwp(v53hD-CMM+1whO#-|iazo1_q8 zWdkN{HS%fmbk9`QQ{(g;Skz~0?P zs(X4jYUVwigV~p|RV@|L5zdTEfCYlFfT(@mkNDK)J$l*tsi3ep=A30R30;+os|<&i zifHh`qdTaH>Nxv!E{l0)!H(mJ`#c06=bz{#+dXO=jf1c(t+AYVryGl~ReK41rUnC~ z-tnJu@ffa(iEHfu`?Imc>Cu}uUJ4Fg2JAw9)pGs?WJsU9Wq8=hlV*27k@=qSZlXL{ zJ89SKjzWyWjJ=;XndGg~T^_}QcPio-q*elndoy2TQGrW!^!&k9ignT3HC5d}u2|x< z%Ib6p>Cs`H144cQfM#*arv%U7F|#{duNJ>vTi(*)&nZ9g6GDQ*`>lN_t324C&v4o+ zot?;1=Q{fD@Zw$o;fA~zJ{L}gb-C*yein#}9#TG=lOWyu&y5GWO;}5}1v!_6k}7;4 zR5lSm>1%M=r)cAE^aWHozInC!X923Z{t#05R!fO96G6q5t6BVwzZWk-a$nMv9$(I zp2^P?Y$3sMBsO^J`J7we)=C#X#sPHzqBj~`-D%^njM~fW0i2$m!YO!%9EJnUCZ8_m zG7oaS@ORMM2-Yem&=RbFkiTj^R?E^Qdje$*r!O2t0v9EVm0IVvvWtd-|ZHRb1B6Q=YEG+ ztf~16ga@!+N_UaM(Bq6V9eVcfyTHEx&K<{3q(gqT`hjNZlu)r*RzIqF7tpE4eK~{?Dw_Q_qo~=s^dACOvp}m@Pf20elqPOo+LOwY)Fy7V zsqkjmzl0pvB8lDe`BS_cOir7`bYgYy2XFANa(aPTH|tqCr#;*%VSk$Lm)Xl!)a$98 zA3=4T?S4*>0{FU#wdu;Vnh8BjWJshSv{;W<*eZuCG49)lIE!1$4xw;BnV9<11RoWW zQ`k^fRu5g#eQouY7U;TC(upmi{wC|Nm-li(}}n&I`TC9)8DOiHv6@{|WoYt^A**i~mOy z#{BQr#r*qv`TUP{G5@}S@xQKm`N#hLmqjoC?A!VOS=9>;4dk(Zt$KlZ3UV9!XT`=p z+jswSIRD32y%0G4-?!rZ?+^NaT;=kAv4lAO&+h(y*pMvqpG*1{U@3^<|DCD#|9+*$ z&;x*T{|18P?3gBhekl9ibp7h#f8dkdz?(?lZ*TOs^=Qx(=wD>I4)(lFGnFxD$-`;G zN85QmcGJ*yH+XyJ1FH0Ze~`L-)HcK_G)K;#eaA_Ifo>?Hzfyhnr`&AQL2vdsQM-R!AwO z(|o`u2>B{lm4L#BiMre0i~7C`wjw*LCy1$;M))0+6@=$*HoU_Np>{SoDCp}GGCOn% z^SATKatH$`^sS+omwquTPiU~fLb;J>v$bUNFQ097T|*RiD@$vJr8zi-hrLC=(VR|> zj=;Nvr6)4ue$kYg5ru7h=>x{l`6X>5GY4J{Iw8_QrUUxKA-D+rb@k3HkvSH3pbZ4T z-v8~KzqU>OsFBVzPUmEf*R)M_5F9MeyUs^6WR4Wb0o!1a0ow@V4FL@hQ#pSI!txZ8 zU!R9Puemku16slN_8nFsri1KEbeQmSWiI>Tr|ob_w19^HWHG zxJ%NtA&Y~(=n`2d?{b_u+b=lWDybgqTj_(|Qe=QYtlew^$P3C#A}NBCdh%cU#5>O2 z^o=sNg#NQ4pWF`dPYIn3>@WjE2i=c_r}7lJp&0WHffrxuoHuMEpBXD~N`M5TU21HT~{vxW%x>0k=a) ze;a`Y2v;gc<&pe>M}&n&3%tHw{JaM|`*FzR-F<82fRsTifNh0J>e&ST`XSnrV)L=T zZM_W8f`oh-)e6x*HfDtPc*fuzvSWO>JaDVH;~j169F=BQo@iE(FKnm3WNV5`O}6ZT zhQ{M0w$)jlXygi~!IJ*Mz$Hm?hMZ7Y{~8b2t8i_377_X@qbrU}Ge?2_L0lH)l}oC^ zq3}ei0oubCs_{eW%y5C zV8%%Yk}!#?0}N$@H!f}6^F7lUkDfcc(8a?mDrcPn6abmm5o?^+g>6XaiJwgkI? zJmc9~*)mqQX6C*yp#*>s06{eA-tM}sE{8Ao{w?KN9^^KKXx=a<7okyViY(7&4)a7? zr3_pcV_6nipR7-;cSyf4yZbp4%C&xTMmck@mu4?2oax|UMjz-**Mn!-Q_`1rR%3p1 zva3q6UzMTtQm3Fw1Vq5#(Ao=XU+!eD#Ms-?Q>j5y{;8t7)La*2V!?+>Uuv6-O3LGC zG?%v_gpoBm2&4wBCY4C{m!;nSKfMg%kQN&H<}S8iXxz%X6`k-wrv zkmUyv(I`AFZ)NU~Aa>K4u!KM_Yxe?P-`T$_V;c{rH`A_R?-IN&8*S@c=xcrigD!2hrNx;KS*YwQc%cC8J!K&p-J?Z$1XwZB6<>GeUeG6- zD6PCx?Jb28F>j-$52sI%^n?Or=mb|Zb|Z9j4ekC?N_JnH_Ddw|edqIs1m#{+paSR% zyLXuPW?Ty<%O-wyNN{&?Pxd9+5xmgtST8#Q*b?n4C(%0qNkYK>AnQ>=C$?e=JJS&o z%Lg-iZ0NWkXaxH&x0rL#?cy?nJG}sRjuA=~d{|ZBJ{D3qmzLlgd)A^2sek}iVoTN! zS-k?0m|8~=jkn$xv;7K)r6aXrJ1iT~3b)YhUYid7$d_Z61xVb00<2ycumXjnr0r=> zR8@Zo9qN--aqri7&SnUqJX>@DeDn$^%ZA(>oKGxkUH6~NQ|Y7SFDSDHK=A9?kVa78 z{2ji94!ps%6^_3Ew!7s@VUzJ3Qu6=Bw5UP^kKdf3w!USOg`1IXQG!eU2 z?WT2_k1xAj_JJTK;Zlvi9Bm6rw4CRPpp?lLm+`w*%11R1(S!v2lYqCWr~2If@+#o7 zY$;8E>!nG_`5%@t%&^I+NA?m*=~tZh^;R{pPRZJjvYqcaf}j*8h}@KPY6S;;$=WaV zx(*#Of1uO}>Bi`T#3ftj7ax2n`^(a{D9y&aMFcm}P#&4hpe<#cQ7^2K<<;2XKfY27K%ZCB0!H|*YEW)ZgOF?OE`n>#@(6nD$ch?T! zq!~Z*inCw@^=OoWM4>rF@AfORiFZAI17_h+tsmgCDf+ask7qMC#2Hbx}ZD7eSg`(-06>prhn9oc(FlF8~unk9de{dLE~X( zOFtBYb`_;aj&G=?hecBeRmAmHDBlj}nu={_C4D+)p2Y-SUj-8Qs_H-bn3xZaCn3k_ z4kz-&q+tc)WK+qHV-gQxE+5M`y}iE`6GkOr5l7e95ksG+1m|NLBIGxW(hkXfSXb7> zYh_+jehL0_%L%3Tb(%6@mIzI69cw%V7xP}`wvZxkA5Yj?X*W#R1HSQIxJ3|-KBYl$0*NA>v*7e9Rq{z>ywVe4YyJ-?q z`7Z(rBF@gf6)#DZ4X`L{(~EX5AX282LyZM1%lt)zN?sXW&B4d=HCZZ%6p3|7sv;*G zB1Mw0<4qPWk##{DR?Wb%=Ac2Ay}g|><7y@>X_Y(AQDEh~LhOD1lKzVs7J7hcLR!;B z9J_+2U@F5_@M0!3r0W(v1-j+YBTS#KC8F}XH)$rP(bG=ijY}|MCafmUWgAruj1DDS zW0{XlCC}xic>IwbTnRNDvkxG6heo#%Y>KeXen=4FWav)Dkh8mNXl6p}ZLX&DJ=+8J zCP+)!yfo8q-O595U|lMr3Apl(FZ_-9m1r7R`ey}b^)5qz=G-H(&H1H}$Z`|Sp$uJz zA-kVSK?1R8fpLyQl_R>GZ0HF%A(crGa_p@$s9k~ZMYxsQ`zo6ir5*K zIc;JRl-H{y{(Q+14o;fr_Fg8RWgA&R0=M)ya;T*u=Z)s2XGvtFc-L1dmDQI~Dj)!b z@{X&SQF8=~4h)%0*>XwLT1-eQJ13 zP(TW!IEnCS(eYU=D9Cs-Yhm4qLe%v)EU)Tq;I_O8)B}CB_}<$#cJkeQ-r|bl{u~L; zb)!rC3%NA+_h@N(Va?+ecp97h9NNw2XTJrRC{O zXoAX4(W99s@@h4Z%MyQ~vBF*8z{0_eXEcE!3ByK|&fVGC4AdE%ywo|7@u_B7LVjE6 zlx3=kRHTwrNyH-qnVb z12Fx`t$&u5oXRYEeEaWP@5~66grM$TdV<5CaZ%Zu4NPglL=hgYSL(^6e{}9m?<5 zA8!05@1ma|=@imKGME$6iH0P(&&`QaPtXKNT&{W-lENAMjUuYmkFL;uIUV*F3%y)0a>Y##D zP?1#WvwX)DpXpHD4M%_|_o7(Joq6QBB_c6NCiknM=aO`q=d@?XYO0(6(U-30aJTv=5vPyURu~>NN-qOVSK_B=>(hWI z=uM1Jda;BmC(vUZm#MIHQp;{IY?H1?mfm(`+Dz_Vle6Em?4U+8Piri;^>uB*@c~^T zDbJX_)gg>Gfsu>%=zP%aO@8*UDO?np=b4eSvl(C0@6jK#2O zZdaX!ohPr%3n!SoT3|*pgZeDkY@V%A+tI@{$@Y2>bx!1&uHHD9D3bG0{f1jToxAsi z*-^1_waGYcuk@%I3JbsM{w?rQNxdW8pkCv5gPXRGJ8`2AVeiPhNjMp3jLdBd5{esNpvL_!oeChaG=}c{}#y7TMo>J-NC;W^n?|fP$HeJabQm|(n zZw=)Sz?_Q_vunfL%AGjg0KjPOGj-9hB|zJ#Z?LWZN@mShOgs_ZAT5y2$%W*HSc}oS z@C!>0*7_;zQ0fAM`Q}91=ZiUKHtL}|9Hu$=VlP?2@82zV$D$^v58pr0#R?nJ_Ap4G zYt4Z?J>d-&n$R-hEk(ACoZ>XifaP2-%lnJaL6fO;8H^g~3TI(2wV*H#!WVE`j;=#lsjPv3EX>&=tx#^Tg6&sCaV^-ltc%GI0X|k z1sMe^B{nlax!KOO0$ic8zrgHRWm6ng98~fQ%|5!VXOR6${o5#it~&YzPgY>+^;cD+ zBmeLbn`S8Vn-b`L@q)Gr^4z8YrkQy1h836bA^;6!(0ynZQPe<~2}RyqkeMJQoMQn9 zM|~;iE~P?xypwl0rj=17_KP%9(dr*ahbXNP!ygM2n4o^FYMZa5pubDTf))}M6WvPV z2Wv1sx4kRJ+t2R?v1>?EC#&kItJS=Gpd40;D73)<$$yd$<#T1T%d3@63P)N&n#2nv z%3|v}##LBf2iF3s^Lo>t37QyFA5Y;czv;$*c4Gzi&~z$m@;r_bff57y3<9kfrtco zL6eXV!R0J_^~GjlM&xYlWsYnZ9`wkBI#SKJuG;*EihLd_>HL6NOXo+H!Iv#N4$G!g zMeq-AMhth8&pqDG`-80Q{W-MkgL@W|!GB(|6GsGoK^u)=fMR9paBbQe_!)e`lNlc* zn3a6@^ZFL!iY3=+TE1_}rKSm0i(H!j<@1$EMXBRpEK*87a20&M()SV}Z2>M2R;C@+ z7w#)hhPdLzM@6VYbY2zHO%|hSs-Hp*+|+bhn~Ne8hy3V|C839%S@EEwKuonm!O+2o zWLoxHLy-4En?Ep9X=~qiu>^4@456n-Yo_E8a=HWp29B_ekV##XMee;?=w2C5LJUUv zr(uY#a1j0i35r5wBZIR4<|&L#Nlnk44$}PwMaMmjppDTO+Uz4 z*0b*tpcBSnRtT{ULR>1YO#~#14XiDZ@LQ>be!>#xLkQlUg4ysCls!c$_w~1j^N4M+_YmoKcw} zqIlt3F5)qxQW)>rmV98=1pdzz5H{`wb|;x|v!qS>GKHhq(!tSO*9DUTwO|%2kqjSx z_>*HwQeAS6GOqj?78m)YIV95V@kR!7r)O{GbXJA}o0J{6RPheJ@c?ob7vosa@Xk_ z4{lIXJ)G)YDP5kjlg$iNn*`bv_|Z)5a_a~Zw-qaL>9Aj6NDIzmuJ2IIvdymmyeMh3 zu)IlV!-ktSPa#b5h^Fk)RpI0(AB}B%f?CnSwMGdl)&bE)66JI7vj4G`C4qei@75vd zn=Y=MBQ{*9oyodQ0fE%ta1Yg;$8+A>en=d2pL=*cp`7Hg$zB}~08P$b{M)P-4f>N7 z`!CEvp>TG$>|O^g-W_Ir$`-kM@nDLO%`M`bc7)z>=nkD4E#&H@eWQb~KidKJho27H z(m`bb?8f%0orQxEYKVkwv=@y`JzlJhXk^>uNp3}VLX{b;H^Ki%AaOPE zMDiv>foDBY_OlH!OTmWr#%1J!r}u_N)dG2A5snRj+l1o>)YXZ+tkDlS zoO$aR$&U~ASb6hr)bHS4kqfk!kRYUZs=xhVF}YiqrZUC=TWacZR+tx5Gt1qGrJ5uv ztnn5KAbRi`y4{T?=h}4#S%P(y`oH112xcp2K}j3{LuvLjyYpdb6b;gzq|x(~qOXPFZQcUnoj@ zZEd(8iv~?-Iw&Kak$qCSrS11ud%k)r{#H7cdB~4pW;V&6tPW`DPtOiZUyg(#M4BOX zo`-geA{|2Z?!kV@q4I0*`Tl&L5>FG~N)evV23r1U((-sxe}z=;r5OP!>$=#B&{6^_ z^eN*-9JY}r#9Tv-JKIA9>m4i@v|ueaFP8O)t~1V3V`l3!S;wR&NFKr8$h)keJI$>U z*$UVee4)a(wIb#4p%LG8K_X51r2!vbO9KqzHyhsQ>c;R$t-@77NfDK=9~xPM{F$6Z zz{${d0E`#D{Jl~L2p9^4N3X=v{qdrxxf&S^)-AX8gB8R5_|m4QaLM|)dQAiuN6M*$ z@3~S-L3Rt0d9v(|pJh(GRWVnTvq((m!(f{g;o9u7=b3<&BQeCs?#7LNdPo%(BguCx z#j~;c3!!Zc)eu@TxzPkp-EMGFDmUo3+dk%>1KMSR={B6bp8){t}s4}}` z&GSx%D8Q(hH51C-wG>X0RzXMKY0gak>L396?~0Gk^bbHpj;uvO5neNCtH_8=fA5lc zAQ{JJJj_V2>q-!_GAkL>od943CSC87HXA&axgf|wA5El#F0tE*a{VhsSm@lWZ(RKP zUEMQ)do9@N$#i*ckTsedHZ9oV@B?KaY~cUFtQ2@JP|RT6(X-&)_l*$Wqm+~(}5c=3BR6PfkCmG8dys6#=ceL{~XPmnBmvKk;od3{pw<@(lk8Y`0m zVhMxGmeR^7K0^wv_Vz=*Yzr(W0V4F1t_H};p3UUj&zCxcud6ru5IRzQzGP19Lu30h zn4k)XRxHoKeJF7d3DU}m!pyCco)e*{i-Yo%MWyUm0Ui;!}VkdkaU_>p`{ie=2CC_@H&zPqpyBonDs@wgy94h-jqD zmsPFbC**gO=k~Z~+&fzMJ6K`iLk~)7eZ3bO3Hs7NR<~mR83wLVJ+B@Od)DXpU-^XpTgvU^PW0qKGBo_33jL-%m5U z5w3Z6V(3yM9Qw=2vqMeO8(C$Gn+@%co2w7}(ryT{>DxcB*S@DlfbZ$YFe3kP&5-gS zj4+d>X6rxKO(He(JKyPzqa7(yAee)NdUy>S1dfSf3wEkIVdu2YGTkJ%_}GK zZoj^P?pA-HnU8I|#SCyCRUMQDkvLe%P*Gl;z_fNUK|WYPm=%bDn(q3N=E-Z97|578 zGFwK@(-{ddf0{vSfH4I|>7~Q0{gWco@Sqnn*Pf_gTBsod1Q{F1I_29$(RWfHK`U_D zBx3oN!s8RS&_2NP#~_eOnH@`86F4}q{t8{@?%AzwcxcmhY^^bT)ctGnxAd)YLO;Vi zlbG+G+xn1Lgz1xnmVnRp2q z_|knO$x)6GGe>`U2xwH7^pAuG5dtHyvGr!l5E6HaxV8u=6P@1s^oD*l69KLBFdi$B z3j#fpJIADo*KD%Apf)t}jt{3dSzQ>n=V;>q^z-PlGwgLM0<+XQbp~Um2 z>Wpyi=BwaOXN!6kOmlgBG8E_MG+29iQn@^1_a**&hmbOsNsJB+$v6$VMWlcdy*s?o zzxQs|>16!e)ud2Ht_g2Ve9F(c6q{@bw`fmgz=|4ee4-hCK)bgC&!MCf)y&|5EvB9B zQ;;E|^yz$K?6pdu-{g@p+FA_jn;OwT-W;fJ?=@&c%T;x+ z*ZjNpT8d`&hLD4Fp%*GX7PRS!)^2!J#okeg#;+Ur(M_u1@PbS352!RrosS_y+>IE2 zfE5(&KD|P;(egej48kMkqY?B=xorjkZBfnAb9s0O45)F|D^>| zqhEXpyhoE5(4-rgK>?1LTWI3b&hU9(6p=PzZZP*ObcI@O9zOaPElK(Y-G5$AT0`s7 z+}^mqA}!hd;*S*YH{Y+oZN9>}WWBmWHVsEN2qLzSags50Wz!#j!mC+T_%3_LA`OWH zba!#kERBLzCNL9t{t0CqqJH}*MjJ!=pSQd5C)c6F+HhYx={s&hH;#qf7%cX~0YLze z`GuwfG{^J2gWZDbHXIp}S~Eah1~KZt(`-H&ZnE7gT7^ZZdfGY4Go_)dF{>#u{kog7 z40h|i6Dyc&@{5ljJ}SkS$>3fR3ev&Mh_B^)t3f3q`vl!-?-$LFHg=5YC zLezCFDax(USCR~ujR>d+s7Sh#Bv~cH)1S4Szv@&uJ+`fgxWgA#;IDYW{ax8fzGUa(TMi?33@pkqTpSET{k4n?nT}erHT=C~4 zSmBAnl5?KPLRv`&F>0=z!!m2_k_;D}s4L{3V!5kg>JC0Fb2?m32i84OBHtJ0T)+B4 z3W~WuLhuUbCqd^M6(Vlv-pm4#je``fIF6m2T+)G*kx`-%gK><$cHb@jX^YwCTb&qV z`Jn26$Dx6JPzT`kc0>P-hQz_T6P?G9jLx)u^DH@D^OD3JRoZR?@+ic+t9g6Bjj-5L zlBQ*orm@xE6IpKl9XVydJ|eBqn@P#o3(ps!mdMyXt{{l%H8o)pzN8&^!W zF3#lZ5$DXC#Od{@$c%oj&?L03%i3;Am74^Y3*u&(j$_e_hmilebKc3%LMtr-M ztF~76#ovqO7*wxPE2C51ZM_ui{RZ;4TDQjiA&=SZET*hlMjN-=udntnN`Izb2jpBm zW0Hh}2IP@+H-VRS>9bk)&M$65Wn4Rk{Gd&Tj)#)A5G1F~I^V@-5l|^=>o`Dkm{0FA z6?kXn?z*l7u||I~1_+DHH0y^Tiox)p#L`0veCql-CxuCT&{RC;qBoD@XR~f&ESvPc z@3(xC_xwH8)$cdiaTR~zx8k1Z*0P#rIT-9Jw!NHX*;}_Qmr?3vj+Pryx&}9Im12xE zl%rIVlE^0t>vO~c$_*XFk`K)U=VC_WoXsIO$*LxKaT=MKo&?%`A0MeWOfR~qAJ(+) zFF?#tyj(1CKIM=)Sm$_B-vI>jFQSJuAOf(5O<(j(KFw}#pTiVT0?b!nC9c2pv%J3; z1_DDJ8BY2QrOLu`zkX}T+FvnIjmt{hX+RBh-6yJTdq^Bh5LA(8;vlt5PZt$JI^jAE z{=~oYJY{Doi^=D=9=$foR&!*b`D<)9&N}4~aqmyxrzJm>`Wc|81&oOAFZ3`i zh;o0x#XXlX|A@O%jR^!VTr|~gLI_@1~6B5Xsoz`lYl=RdGseTx8afW4p@dtGz zKn4?z*F?k%^hw}lvxvpXvGci5C(ZJ|0V#o6pC#im<;JWIbIOu##>FX{Ry?2pmt_!< zLMw-P1FecTS07p4@W-b_Jr#&B;zHTV%LLFH5;*#iO)+rf{U7wa{*KG-v56#UvyxSW zIv3u_zkACCSEsl4PjOElon*DuBejTeRHjP;@+O@f7xbq{5O$sIbeGhVp`c}q-*Kca z)iF5Kb<~Aps{dUjlfn_=TMpogY<@!xNotLhRnt#Wh#jZ|T8@R5(TDzlD$2?dc-oR6 z>}OWo{P3v(suI3EU9Cvb?W>;iW|-g&rqF|Hg~4s}E1L$Ws^yb1K+J%&R*)|1_vJa| zN;0dGpgyA#k7DPNDFFR(&q~>Xf7s&)K>78fu3VP+2{c&0qO`x*ufZ~QSzXo3WEH2c z?p-c~0Uh|{YloUC9}Y*)V-KsR(9&aXCVy8Ei9?gz_wDF^oqhBN>EyJ0oWaRHu-!1D zv!!|a7dK9~uZg1UawjX$7=IV}Z`4`S0Z|6cpZV$jVV?p9%YwQ{Hh*@MN_f={SY@r`K9P~~@<}lpU`O4cTF1Y%L8Rf%p`=AUx%YLVm&`TYzFEaB{ z^{Ry`B(HmTWN{FJ6R1Dn7Cl((_{(>o( zBxHZ^Xvj;$b0fpLqS`{CVd|+|hK4<<*m>FE#>_D!h?wdc@i|#P)98Vpr_Ps(JHrfj ztYsO7)CD`l7Kpuyl_(a_wJd>g_>wMy%AdHZHstcJFmvXB3IqTn$60`_gS-s+=cP?^ z<6?&@$HD7V2HRbm#Huv6e`Umt{8Bi8X!O~mHg_wkAHdcfbBU~j*(S?>lZy=KWA)jm zJMi>vHSvxB_zGXP^sr>hD1R}RHiodjOa`q8`3e4}NlT8J1@)uhea@YH!%Qy@K@SQQ z{T>S5(^dMF#uK$BwG>^mkLO2!Qu4$Z0TJIlc-4+zanTA4;5MGkg+kE|f^R4$;!bPSLo0kKKLE zYRlTF$kOf0>BC$7sGrE-G1oHhY#uCe1N%vH2))1mo$R!?XxP4!TBQgQivKW@9=JjbOe4{_y{WH?lF)pnnE z6MfDpm{go3>+y1VHT3nkdt{GG>O&dwdCQV31%`mSQgDzmKil{oE#z2g`$sKG`aB3! z3i3oMg;$gNKs$jEwCKzg;MLjtc6#zYxEG~Tub(7sD*%lx4`P30Q)CcAr^h&En8$<0 z0l(KbjgKj2wJ(d@lzeBXe@xR~`eRhqqE9}$Op0l;bojAhtkWpcbT35)4O6almvzms zI6}qjZOO7G5wOHS_{@6AOytVea|T!>@n8`T#Va^_;of|>mdcSC@7yK3A-Z!sR zf#5ARmpAQbDUW&DCn^sOax9Fj2|lPY64^%3SI5}hko$e2$%ZPNi=qOsEwAxBJ?y8} z|1qG@rtE-P+6`_}^z~tstyNCfXBbTm4&W)uavW@xg9eHoNSr%Z4jE2js*PYeo@t7N z>@W)0NK|UJHgir#rmE(+pcu#~&9Rn#Lu*wOxspXSLX8%rK7lz;#&D1 z^!a|DS2UM)!Y*>j$=OrgzuF`ageR*cA2p|2xo=Wu0J@dyvCaTnW0wahjbl54wEjIP zRtY<;3T?vMgf1~NOYEgF0*LgJMHNojQS&1t?-deOE2=C`DX4fjl2oO^PDt6!jR&*8 zJWrE#BO?bl$6c`v$WNGhZ1JNONwZAV!=%eM22lWSF!$_WK>Zs9YU#uQ@B7G>nCC>= z|9eRFIrV#v+R4}umlHXT6evCPZ`y+eM*x(sb@r*_h!V9fr^j7WPmY{hmj}I^AG~H} zFW2};#?yD+h#Hcm^0PhcpA~+MzYw<-JI#+IX_%n?ySKf@FYa`8XRLC#pZnfN-9F{h zr~VyHM@paMs*+RaO-9(Dl(iSZuJ-N%8?jlQzH>CjI;^$(9}N;zmSHqH+u1>uo3}Eh zd}g3ChRkSWE!qB72B-%i#cD_4FK_D}2p9V^t7IBN#?qG# z-(<3Lpm^Us)Ow?-^y9;l8@`Xa|ER~1OKhI2PDT)Vt*s_3vH6s-k3d+Se5V)7uKD!c z*AhjRQ2+ZidU(lPlWOA9T8+BraD60;I2z{@qH(YT-iDAovgu=Qd}J^{=9FPwBtF?| z-r%5;Py6orfQF_nVr@BU5GO$*yZVuifz@5eJiw9eQ3btS(K(Nv?JHZ5r2;F^4$Iv@ zC*AKX?@hfpDLK}_`bEZVQ`h7duww6$B};GBMLS&5k}M?pXD^-o_SifuHZ_{B>|dny zY9h|Hyyh&D-OquPn>N05eMb8---%&Y727>$Z%@wadUAPiyQP@?hc9cB{aZZ@-0Ew3`a1@?q2l59@a=M=K)A&IvZ<^AEMK-)AP>1KkZME-sy4{CUDJv&cXkxbh} zp4gLITm$BXU(DKWqF?_!nFJ}k_?Q_g`I^+@5KfsNOL4E=$b0o|I%b_X$qw2mWVuw)T`*TF-dbjfy>t{=Dn2qdWpz??p!{Nlf?I|Q%L8Mft2rb z0<8q~@r-siFNgN{ZuRzT6k9raZ1b2o9=ihFM>5+W|IE@aveZtNt+Gd^->Rq5g~Fv| zIl9aWS{E^wK|lYsmc z3LQ+FFSpf)VR)^pphE?Mw3X@^ggS?Oe8^fCqxYz^4Sa62%U)mlpycO?yDp=h*8b-HxKMu7+Wr@G_Kdf0hQWXxM?fuO_6?F}TAt1CBs$rf zeE4GKg`q<9K0~k`!hvy`B(O(vLysa&5vkogbWHKYN#1f)$CeQ3ez@ux-PfIu$J<>L z3l=p7E-t#N@b%sfHBoP-i)@prx*BIO<5zGgy=tM+2I$nIZ9odvCSz^jM6_o!?iBXt ztNgl-Y{`9c)_29}Y@a{)ZEhwtfk063y732i&Od z-CiWyWkwYsoZBq!Qc&}k3fkQ-{BR{(%?bhb{Pl>s`blEZ+zuGxHK?YhQmMj7#jU&H z_Waoij^w9X6}-}eEW^8 zSsz{BAkLS%`YAcyOdC?EL3a27# zIsjl3TPx_tW4aAwA-P)%b*DXJY3&U=(dB!NN&$5*Tlldtn21{J_&V7qW_|i+0kN~Z zb}hd)Vh35vB(k3~`P}akKV*uIpv&@H9lv|IWM;}^W;TAeC!xCOg@kZJ zb^`nbH`Cx5at^2a`n(knv`L$LYcgn#Fzyo)KIZXLZZR4{<1)EXZMYk(2X(9oGZ~N} zy4m0S>+0NS!Qnz|L^><%SY?zBJ^^}D#C;dRNHZeK0#@bd>i65sfy$Ivu*j?go{cQ; zn9$1JtP{R3cNK$9XAVn=PFUvkNhgyr(Cy7*r90i|`T@sr_r&!B{f06|shPV#B(xKn zr$^sA+R^u9)QEcYurk!b5VNS`z1@S#Trx8?XPNFov2*1(i0+@7&-SVsrsVYfY*b;k z<-Xx|tT!q-XM&%;&q>u-07ok$74?*9zrd2Pz9Fz1i7Ge5Re&@pFM9v#&@51#w10(a zTHYtQ^2|~cQ{%(N0{nadoo6QL0*P-(%M3e?i6?u%2?q|kx|wxZN7y@1P9G^1Ie%`= zA08tW)C~2Cyqgt=Y+6F1fqBz_-C0Kd?@(Bl@WO+5@J_%U2KxqUg`6nItbl@(wy?tb)1}n_9^80tJ=lV7+RF?yBM`s4 zddg4MM%rkq&b%+pq5A_7HdFy8oI}-#rbK7NXlXhCB3;c^?Qf#k|Qw69IqV(7?z@6RtM7JZttFDo3 z{i3xq$ggK8HFtTCxmX}ot9f$_SqsKnn30WgUHUvLQ)f-OB#uqdWHW<~kEWg|!ect{ zVOAXW&;2>Pte=mQvK2Acve}P<*Se~s2@i6RE%;2OurKKP?l@s;P@`Rj63C0hcnBlu zq&!l6L6iPiZz11QWO-d zu6hcvkPZ+YqWo}81sOPQDGO)Ya$Ghu^Bm@o0uyAqP9qmk4Xr9uy}5AjN!d&-WaFQa zu$`z^P@dDu(qXiF;-zNK?b4(~4S~J4HNWmWJ+;M3VQkn89!FoX6aT8w0aq@Xi>3UZr!<5#;!5;+%0#2Gu1k+?|{h}bv<&ty8Z*2^H{y&^u;5JQ%1m4PO_`=MF8wwu~3$M!xdq%A>mlvk!W9B}>r2 zSBkz(%HmJ$&Q&E5@?8W1d zczC1@zc1A=zM}9~C5$q9({0i%^CQ`fu5>5iGPJ-YFgDUg6!*RZEK|K(4jV)vwMj`r zevpOp6`QKCO1p+YlGTH(%Z+h`P7bHPwq4xT5Fppin>TLl{RX(BjU%%~Lom-G8+>+r2$|cJ;vl?m z6qTx%OSpSVQ>u7xAy3GDVeku*450ypq)0RR?1TBluk0h@T!yR8yQj$+uONG7N`qM| z^TrrCvL8dh+!!5R_}*_TM`(R68Q?y?i`+kcg%La1}X*^wEOb(6dJ{tB6+KPZC@A{ubMvvE?x z*zGjY$2Z~-IWw-GsQX#t;J5D7{w@?^FL`p<9}l9{BmB0JI%(U%rqLQ>hr9(!zM?+G zBeds9_>gvyVUe{>MM9F-+-dR#Dy{rQv;lI95pKcXee!TI}fBcB>SBTDuD95}Z zT}Lo(#y|2Qy-Or<#YtiW-)#4cNG2hMzif4O-nd_ev&nd8=88!waIB>%xSNqEtbO56 ze!XcxH$*6lEhDlH{WSV^NQ;Zb#cu~zAO-ifx|KD$ETb}D(}oL@-l4E+8(4{`hH16z ztl~@p;nJYg{>qzI!&czLi;KD*BQ4V$ztuwXWik$POGxEF@u))RkSTsZb=ZrqR0%ov zuS+)$;w3n7O}GOoyDB;=aM2-}ERHR8)gfG&7RZgMC+klQdQZcviQwg#E@=yz+BJQ{+&N#lsqYN8 zWY_y!YTrxB#9cs^1gotnGqWiA4s-{n10(~(b~DNTed(dCne|`=V&tWaG2Ml+ z*#?&@ca}XR?3tYV)cLQT(*fQ(M9pSJjVM#qW2>%l?h7e%)A3XPk>Wo=YGyBG+pDhuqF<1s zO58bvQ|DlJqaYx`k$V)r5ozbsS?{s0E?(7&ms%_(wP~=Ns$}L0pWjn`k(aYPBV_} zZ(r+IBPEJR>KTr`g1C|?YK_Y`FKPGQ1IuR!PmI#y_;4FdJ zIZo6m7Z0Lh#7P9rs>js59^bI_)>{e-1p5uVJ6bBw%V(iS&L7{c711f@ouVmQ5RODu ze}pE8m_%_$rKClP=u^ayPl-i1%xRZ-E`7)_02+9awOFf)@zxS#xY}$iSC)fx z+m8_VnG`?{DRn+?Fq>RszjA~acpN*M&tdV7JH779ac1cbYO`G; z%)j!%5dGyjrw#*mKEGO*itvVc;d%) zI^iW%9KAKsOFD(4$srsVsV873;Hdxr>@(=G*@u{KQ+UR5K4d}Qe|zFTWE4KVEPaCf ziC3lY&dGQhgfLoqme9D(q9Eqk@C=<|paM>XlS4vf2qN%B8!T6HHyMZ8Pu zqpQw$-~48`BV8wPqe#1B&n+U=bx4C02ZLFC&62hmJ(V%&r_N-N#-So0U1$IxQQ5U$ zt!P(YCh&8KM8xJZqroGUm*nbXc5@F>b*`m?OS)EKm|Cl)9fEvmt6+}w#z#@VCo}Og zyFYng*9+OMrs(~vsd$t;K2mfp*~R<*G5+64x!F$qBULu5lZ4sxVWvv%AM)(3zns`o zrCP~EppoGVT0Vl>M=^<)c6*Ap=ZqWRKtgBka}ulJ5G+#xA@=K^e7YTgt^-1fthOrD zO1RZbougm=<`zJlF;ZFlY1(VHOdm4)Smm@Ct#7_e?PHlGfC=xlCcQl}yCzo|oX~gn zM0BWv1&2~OoZ$K|6vd%5L65W-Rfv`p}s+mv4v3=B%X!(CMD?F+XWEzEU zCK}o@Vbso|AN;PLtX^;h zGH)1w?Yd+0F%(dNSpE@q)Sc)PogvnT`RC#(YQDm?XQ`Q`T;O@tE*peuV~XkrUxAK# z0&3ak^P(VJV{p7U0T1;0J95$b2r2ID^>d=4)<0;G8=R9!{j)vpr_z_br(%~+()%r| zpk+;2vNG3DjVmXKj@tkdM-f*OW;}7>DE_Fv_t@wDvAZS_35y_PGDG z==4;ldLd$TDU~6@AW8^7>U9&_Sn$G>`zW<^Iq>Hw`~IHE&+W8kFTu*M%a03Ti^xpC-JO+@7u;p#HlE7kELk0XP`cC!Qe5gcW&~Gz}uU* zB#oZTR0$@81DFuHRjf%GQLO0>m!wAMvyfjkF$rwJ&`){cW)E9ZptyD6pP&OsRLICn=;>XMC{wn-eFd9eK96b3)8!Tu zaI~2~tB<)~d@st+XK%?5V8CBS;ov(CB<*{wg_jp2fbWtf*-lKD8tb5JfjMN5b`kUU z>PHADA5~WRmC8r{ir$vTNg!!4q$CiJ9F%2PJyu7dw80q8(27`51R*PGOaougI%%>!vVW{2MPQ{e0tpF+~T_@dA$LSboTG&|6{B|vplM<5s z`?VF zyX~M_PX|!r-WC#^pZG7nVsbUwOmrUZ_UY5j7Gw9hoDC59J5(60?lQgiWW~xX^G+|h za(fwjGu^YaQFWyNWj-@1_N-9mfn6~8s^5@>LJpZR-decy-n)1D6U&WyxicrpKt@<+ z;!PELSgsm6`Tm9DGH}m|w@lw5An#adw9GbKAoWCJvukj#sS2y z&qUxG8pxAhvcJlR?voW{MO6#WFD$Y5x8%w}`3O1>ctmdReBk#&K+_3(#IPCyqX8#m+B`nS&$<|v`}5jJBM_BUXm zW4XR|jNNfU@#^(=)q!wHaxwb6+Ci3Ej2=2|&v#f9A(Vk1=i1oNTe)MOL63&SY4Pv* zSUo9N{>~8u&%I~s3B?qDaRE>q!jIcJN@N*XV^@b>U0SjM5z16!%X)$-1lPq2*;ZS6 z*=k?ym&^*0$&mdq6Km$>23jc#AtIH3WJAZ@(L;Y_&OYU8SjO`Z?PvdSgnoq$8lVoX+qM?uge5lKMGQ~c5^7KntD$oJCx&cPwxGn1t0iY&V;@=;IU{RmbqHgO<{Q62 zDu`LhWq(R6c^+;N`53r`Z=>lUV6aGs|T++>LKJ|1&g}k1jKdsqIM{KlFRgWmLSBbmrn}S+z1b}E?J0V6InR= zX^>m|)O7lpG~&v#H#})NYn&MS(wb~~NYFBC6OI9Q8)2&{zzB8+pUN>EA|LYKkM1V4 zkTKY6jQS zE)#UZMzDdSzb=$Uz;&?evGB8*$+Z^JhE!I}cIowUrO8T*ar)&8u;*-RYt8_AX4^sF zIDd+B=uM){HD*>~MxzMXFFiR(o4WCLxPE{LEa=WwOX==yP0ZFs&mFQ770Fe{3UNlg zPz2c}x$h4fOqRKguDApM>tw?^iTC{(CfV&#q27=l)6d6!I7(X#$6v!?^gNL|?X znc`n$fG`Ccf@(EBJ~%v!^?bY+ev20TYOmSfTY*qpoEcQKjgf%Sk?rlx!gdk3R2@~Y}}pvCvk%tR1R4_T6a4r`DdO8sV-b-W5H zu6vDQKvxbc+GRj#>it6BzEq5roNw#kbFs-~w(~MzZX{>wnt+9ecpgrO0J+`$LOv@+ zBZ4c<>s>(dr;I0mq!;(r#?mjA9FQ7Va&9% zkT;OkDuTAl8){Y15n0_>)DsQRuF(vu@Z{pH&u$3pcva{sT(5o_=Pn} zQ@Dz7TJpICmw&^&0*x`aiH`(YEvr3L{gAGC^t#$V|=Z9ZZ8BO*){H_Md!=AB0 z4QN>~cmUFo4Jp%NCtx=RyQH}ZoMg4wl{*!RaaeZb2cCvdo$DJiz+ktG$0o+~;PAnv zA0bO20!+>=sUUz`tQY#gJ_Ig~g%&jz^y0e~zA*75!o>ThV~W5T$p%LakeZx1?`UUp zJHaxn^(08AV?zCus85ynI?I2kDad~qzj81Sw@nwDZc_4~Rfg+9ct<0G<{a#%gE!`z zTc3NCU{qwlW}ezN4Z=p^HPn7wPld77#}SKTYM5-ITkS+u{blSp2ZP~{C;A{TaOIP^ZI`(cwZ4fo@OWZ%-v=BYzOi2k*?_dnIsOa2f0j@ z4#BhBO8VA?Ixv|FX%s@Np0&|Tgl1!Cam>(DXA(lo?upv2T$@Q&-oeT_bQv)IKHH9^ zA@K(tyk`o2td$WrL%?wa^9 zW7f+OqXr>p{X1*#Gn;`Jtzjn17l9=nkmQvN4)zOe%;gT9`a zdVyZzH{=*mIFei$W^1q`*?w!rU8vW<4S7LcJl{@`kJ!NIL0U73kv@J?IB0tgUNa!& z`g6Yz0JMA?zx&@7#zOTjqzabu1I7l$&d|~Fq?hkq_fb7R2zCe2qHaEk3+u3;!$2C1||wTDbYm3 zUW2!nyZlH`_ODdE8>99O1ki>7hV4vkv9o2iF9W?|`!+m=l!qVFNvEfrs0qUwql(7; zR3oO|8x5zjJ~%b7rZaTBpAb^{BF_VG4Xd4s0%3g;ZMnAW1IZfv`s+cPDUIJ~_lHG| z3x`LI3e}89CkP-v@Pg6)wchtA_`|7~zpI21rnr9+9f*_9P8Wtg8D%o}KZ60hc!f|3 z%xD6n5cU%W?%Rn>Qf)qgv)U$qBIifIZu&H~Gx-A#8IZ(Qi8gFIt=nFz9v->WTLESm z=S&b5IgQbTHouK}MGu`OmR^qSK4tAKAS=cmYIO~GgaE*TA)L?Rg9MWii934ByM$sk zg-Y=dc!n0esgusn%#MmDfTld2V%hT`owhZ`QYX{+0<>8}ZVu1;Nw`P}=jFF`c@X-r zs#7PjtK;Cocam0eF-17ebOZD2_eSQ>VOce>_o%c{e-}D!$ji>-)-rFYMv4H8Onmnw z?$;KeC-%Oz-|JHtz7BSlB=on=-#mVTcR-{-?-h-rN?M*ngrdm#xQS2B#XwarnHzWY z1!K?x2D4pR<`Y}XK$=zfD^~v-gd2=aLsPCHJvHo0X(R(%=q7DWu9y@Zp2q1C0-sC= zOX}%28mwaYZgFW-Yt2utDfX!wdXSvk$7yOgv-mVDzhwm|?10*cfXlLj&QPLMAgNt9 zumj*x2S@67gkf|B+!P(P4`c3Jq7JW<~r_C#j=Bq)ZgnKorbvx>p!2V_E)qR-+}#DZB{8? z!I2p;h~-Ak6h#DVj(2EybBaCf2s2T>Xwr?!#WXt)Q(4g@;0cZ;1;a${{F4p~i(lD9 z%lG}rFqVkJsC`HYF3ZASh=h#L`n975Zai@RyWT zfI4fsjkXe9NrD-lxpv&SDWpU?@$I?L6p$c4?rE&5qC=R1S$Gij+%{0CM5ea zFj`N{kL{26Kt|>dVBMW0o8m$pUH=BXP-Z^ghr{cBjGJ`^LSwXBHgb>Aqvg5MW&T)% zY+tbuR4RV}O#}9$KVUVx@2SFy-_M*w4hb=WodnKE$1u={?tjj|Pu-$H6(DA!GFR;I z(=eC9EI95G^aD)Ae(Z|wjo~G`ECf`i0B@}#>&6R7tRy}U9*HLIz^1%Dd0W>#R1CQ<=`#q| zH$(QgC9$CA{Ak-PTuSCtGh8`8S5se}FbI~$)0#ob)L$ZeY;`l%0Cv5{t+(7OGeKmz zMX^yGt z-w0*8!vv45XsX(tosI5fz|4xt1J|G-5zBlH!`X#+x}xtx4z{7b_w^?e-=q5IyH<(! z(TkY}>f=`^t2RGG7;!g1A9M=&9jSLfXY2K_OK6Ig+XGsxr8AT>^Sl-fnJ=?&ba7?m zU~hNyJI30BrkXL@j(8#~DteFNC^iHCQpV4C*{S60KfQFjGeEzhv^+Si@X~w$OtNG6K5@tFYoZ1O zQ)E(LRCBSR2Ouyl!G3XM7J7*|5~yX`Kx1%qhXVBClnfCa;sZm+!4zuO>tt)PxS=%+ zI*2*EETsMC`-Z{`VcMy8t_IV#R)*9lLC(X4bb;g8N$czHGdB<`C>YOzJ7l|8s8hn8 zlI!2}q#-|Fp}iR!@+_@W1^fQAxboy``O4r|lv=nroQS$z(I zV<_pL-FFLDa*@9{Jik!^;$9Bg`q7QGzg%}&1Ln7-(tiiqg&6?elX~5xr$25GUxg{A zgUhbccJy|i(XswMPLTa{PU15S)?I)+5fbNaZY(Bv&UjyK5s@mq&09C!mpxP+aJ+xF z2@clozq=9E#7cxjppv z4NSUkd`-)s~>fhQ8-WE@~Bg z6=Gwng4s3!SN&$lxMkdcjAz^w8@8F>nPsdy4Sf|pL+>_&Nw$=wKq&{1osSsd3Cl=d zeL{BA*M6jKyBz+eF_@}wOd~Cbt|!wy_7Ji1=_;*YII_c`^*^}2H%BQVA^qZS^L&M0 zDmwbC@82>!wpVlz5^%?cRVYeBDKsiPX-nAh@#nmD`Ek=f$=7@yxba!ZFVY{n`!oMT z2D)SUO#%1!GTqQx5Nc5?X2kYSim>=%7g3#1B=8|wpT+b&jlcGZ)ousRhF4aOEC?n; z`UK$>xSMO=^|SIEwjsF?nVX({LW*(6;=)a;2$QYxLgI;vF9m7dIw}Y^8Xu6X;O_0K z4v&NW>4o++l~?E>aA*4y&jx)Y*FXSI!4pK?ULnfgpcN(gy$xq)@~4UoM;mtlPANA5 z2DPd{R>CQT>;?`kolWvr61qUbYd_w@9es=t^U3z;mgo`Tk7T~~MX$tb&TDo8)%Sv~ zwi!U}E@UaP)ajpMYX?3xCww~fVmxT-MrsMr^j_&V92bXF-1*yrMG3|g-Mb7e8vC~T znQZCX_LooCg3i!&y9ybagn5Sd2Tetc|HLou;zkA%4~fnfO#-TU#bvO_ew#W^?C0F7 z4X#$xV$j{w$O5X$KOzms-%l5xGwjCRN($-p=g&Ak|E8;dXprg`&2WFwg*5EvVKAKX z$&bd1>{W7$@(I>5ZDoNgCj(sckNgwx2S%00W{+zRbK2r6$c$A@rMDs$@p&jCH{+pL(Q=D)Y=<=oj>4M* zf@PGWt|T@jh=frArituAVWg_B?H&}c_MSsS4)Q3XtoemunLPeykXKSqb)6E6H| zON}JU**WA~d@68&ScxC=VQdalZ8J*yc>STYZS0_M3kx!CpZ*01)D~t3AlXTrA})qI zQLGB@>0RpN)JOz|+n0>y;Nr{VCG2`iy|z;k{*`3I%@ zBu_g(AfR&UL}*LF@eJG0_VSqup4s}zfcZ;g__54MlDG|AW@eeCNnclD z0Q6TEHev~LGR%rHyE=acYm&NEvIi{n7{^*6U-se2D4>deA0-90l5-kEW$}Hb05wow zqKH6$3Jl9I#j{BIag^RabfvqrtpUCU@vNAOexM6?Vg;q!dnr$FrKtwU$NX8%dwH<5 zcMN2x0%AT%60;K@$hR;wSW)mMBOgrXQE=;+QBi0)an4h^eLExr$>a>-_hm(9XTCTw zF8Ym@2955wfdch%$(Azz_0E^xYD7hnkld;}jRMnqbME=Mgzw|sq_~iQkCFSQ;M%b! z%;djq*@cSPl8`d|^K840KC2DJfY{vx!ROAONAu0^sck&|vhCBz1$`%WQ z=4mOv1R3!3c*F8uPQ;jQ%X*VCJqkG3@XTOM?%}aeX%9ItX!@RjkNJ6A*mI1mTX_S4eOt}xdQ=Sk)6N;0zxWZk*%X=&iUyFb;GZOhB zQsR?ip5=5R5Hw9<%#sdSS9B3>VXEH9l157?_Q#6;2yK{$Xm|b%E0rKKu2cR=uPBWo<#{q>;GrG1R;LOHL%C6=Ojgi zl-OSr3~@^NdrCB@wzaXj#qE#7;F%Z<&#XTQ&1OJZ7V;F5n|Lrz9%T?ocUgIQhwq)do#bl`bAFq40m^tM9tW4k1$ z-i6IOI~BhL5bJb-n`@jTNL2xPlS9|=*Kt_0Xj>e(@TU;bDm4>^irfvLbzgk#1)tm! zyNm@-CX+t%K5l}Q(1Jt~xi>*GJU^s)_*4svh{8MbRi93VY@$Bh6j z9Kn$?Jarn2r@b})*)OhhG{S7Ht|kYykGvn-hH8z4;&Vu^1rBJZ3m7MKC3jK9ft8FS z{J}B&$Dd3I^t9{Z%?SQFptH$*vT=bgZSQW~$W>ezR#^G-en~I%=eE+vBVyIh0}TM? z6^^5*s4#`SJy53wFs}9M@`QAwU{-?4Bbzq*8>;+vI)2pUUKlBkS-VeD1%K%I=C;|^bkPDLY*8U`DIN>w?Ic38u4kAbt=jF~t zM2;5X1I?Kx-JlDe0a3$oR)3eka+FUpJNCLf{me37&GJKmwUhXQT^CkPB%pM*{daGd z>6vO=(%LQ^=!=NhG}y(l1h=C(ep@w;#XYHI@gZG}dCw;|{sN4M_*gJ3CD@sF# za1|w-DhDZk7%{h!~PkZScFHWSqV3F18uU|bP$AELSQ;p|Hu=x{l* z{>w zy7T~1DGzS!-SzkQ!Yw-Flt&-3KXWo2iiZraooI!EyI8S*4_rvej6u5I)R?pglu2$u z`b7A40jlmE(TdyWcU@QBx^ZR)?(ioAS+6bTzG>*ZS^%=v4W~6#AoH&jyJKe zo~IidoosPG(}$hYGGz4-ZTnz_Do1?=S`EMf!`NOOiSLDgi^!@(LwV&l3_QDbC{slc zgwEq13mmbi9v9%X6bzJ#ke?fi-iN5U2VAaI>Ai=n!gGUeHI-e^t~J9RH+q4$TXjMc zRM^I++Eq@oy=XO1;7Jxv9vs6V)2^hq&hwFt74DW+yn`Mo3>Vmwfe&qdW*z$O5BqC! zy6NBO5s`d{|9hnrd0-esv}EhC5{LHL0ebCimVCsg9rr~AZ@9A@Wdmo0^tKfS{lpXgcyHA zgcKi#4T@>$q`6W|I+ai@iE;BISryH^tkYvIXXo%2=j=w1KC z`zDHy(^0%ZCS(r{5S_;-FN2JRZQ^J@7@{?*4-qRC!c?IW#N`S$@N~DQeI-YU^JO1Q zK=&&ot01F|y-GTHtmO@kVgS}=xz)QS|$5s;T{lL)&%&kUp!&SdPd!>{U?elm3HECLxmTBbc+X34O> zJgKB-N9Mb`^(v_oDX*t{V0XdV+HNTT7)QpoS*HQcHk8QPf75(N2}X10U{y`^rGtb) z(><$hq%h+Y;$@^w#M}y2#xaFCG6{e~Eg_cKK03PIN{szdfLC)rw2l>lQnW2gVkY0F^YB`C2 z{N>E0H1m-vA&vj(s&F0exYU07ldQEEZQ&4!|LP2#MElWiOg>6XE#x~ExpK~BE(ec| z;WO7i!!|=k!j(k#MON|aD>7YOdna5TDsT1Tj>yO##Y(+Oz^q3Z9M^ScpI(Z@9%h)H zeMVqsObB^Z;_2+F0a<+&Rr#4GSezk}Rl>pm1!$BZcJkmE5dM#I$RC^Rh0plnjPq1r zaRQgL=PYb}Hh!2byP*ER9LQOxun5a3{F19!Jqc2u27E2UC7tvL!S~h8m#xsQU|3U0 zv-$2SpED*lD`2A#ob5$|O3-E-baVZp&IL=76nQw7rOH$`{+x+mNee$Rn`N1Gd*SXG zaOOSPu#)!ZF!yJH^AnD!`e@hR02h}h(h<*OS>Yx)kpbO4T_x)g;#5rJNN7l-9GLkU zFp>w851Mq?2_;ljR>Y$TQzR90<`K%D84cxsr02*@C?-b}ybZzFg-pE)l->84p{w6y zMzo3hM;Uw&9mpcm=byK9rS!Qn1>!d@XTb(jK&RpHNM;lzLI(;mc{z z4VQ*=Ge2KIUxmg#GQvoMr9;AQWkjtm<2#WijlxlZUz9dz-wG56V@@i4TXvt^*FkF5 zhNYHPlJp3Ai_1Zt7)u$UPbD#z=GMfkV3SVG+`d@wi+LzOX!8pXTl&k!I>g0ka`tZ^ zheDz3Xna>7WfN9YG31KziFH5ANb#a*;Q9-hS@L67nIiP-JB-@5ewH|Jy<>#A67+Wt zW78m=EY2;L>$-Yae0Hp_yw0kvk1{1$Y0Fi8TDDK^R-ZTDc--}1{sfa7(w~rie(;t_ z@)C!Cl83>5&MO*E#42GkaTG(_7o_w|@iws2>arxL0&IEuMwX?+t=zsZZTY8?ZZ;C) z$s|oKZ}dS;fKMBe>h#Fl;QbD~xV9&o)HXPnS9V>`3r|Lr1^%xIk$&oUHXYaW&x#Q% zk-R5du3Q}=Mn%jEkPqCGEGPJI{bxhyr@evEg!J-uHeU?u%%o8&rd`VCEsG>44yJiP8}`*3_uUi2h`=D5J5UPA*mPt$ER&#uld=*h z7Pj0c<3i@S#GjA3wsJ@@bFCvkCe*W=?dHHbH4@(J^&ulLCV{WAuaHA)eALC}Wi)TM z3`HFvx+9r&q?Jj^iCx+ z`RAp8cUcuZES11H861?G$|bTSMo`zKlEo=S<$0m8r_$}u@AOgz3i(49E+cxUvRQ&h zo*aJtf;21Xoft^_I|rc(>she!nO3R|VoE7%_I`R?$Qh93lfNIEY(?y&BF#v&R+Th( z81dD|e0NoBACHdpllJAP`hNH(Ze?QL}$3a%lA*bRBafazh`Qw%%MZwm8 zD;CEcGeC9b@}&|L35_}4%%G$xz)mXWGIBr{ghTl0_}7WATf1vrWSy|x8M+Q`O75C6 zlPE+fmxO7U#^|D!7N{c`li*4L19(@?C6C3jVNQ9_iAmrK87PK~K=Q46!!yP5J;ZN< z4+wap%}-cge8zsCNz0!Xt_;E=wykI<`98PmyuJ=4YwVauLmbHF{-8^*Br*8YV0Yn- zxc8+AnD1|`9iBBL8=gI+E4JoOqho_BwTxt7-~vJi>Dq`&#Wi8kw7b0MJO1XXk0WS* z_l$+BgBW0<)sA+O<*!;0U^5EPT;EHz zH@R|ask}d6V~7eAaT7qJRBnApGn@zImh}bz+F_JPGvVkYDXt#=%KayrlmfR?yuDvFqROgBPeCI$ zUnSd)J4y;*$(CRJW^T}*`p<%$~k^x6|;$a&^V z8Ay;bnST_Sv@X+in{e`)o82#WvwFa4V(8ro3+t%>kgS6n!2FO>SrQtYNAU@5lfk+|4DzDqOAiXX!-W;Mw)$cBviCGz13QO-Zu}n?WX7hERjc zR%`_Iv8BLz!G^mNcQc~rw=wGm@RZB;B)G;T9|NfZ@&o&!2w@mp!?S?-;7^9-GUMC? ztFZ_n&J%C_$U||GhX1RjeZch9o6z>=tsTPQk8(MNjRAF-+>Bw4W)_=N&vfx*VCX)X zJ{hLld)2vMjS!&^mt53Nc>GB5uZY_Nh(Odp*UvK0&jB+(He@nLsID9s#B+6C!!hM~ z4oJ()Ml&D-(G&7?Dl~ZdeBxP@!z%VaGa+I-GF-nT=nj>P@Qc#qIMUshir_qRC1b8q z4yk7_T}X{kbOVzH!l2yYx!AOPk>QrT0le(7ur}tnP{Y-eSGv!gmhUV%Wnu?fkHX)V zA@>GVQ<{QOZV~>RI(KXb88-75J^{9UUI5xAd1567&-emJMDQb`)JSK^+#$XkAXo6b zN`IIxvJk2m`%;mMwvHg0$Wr{gHW1$~&wiI3M5?jR{L*XNetUpO+Xt?fO;~>R%q4xk zkc3gtc5@~!zfc)!(p!Kb|a*a6}`hj}-%S@7Y) z@O^rhBnW%v&WuHk@=Vdo6loT`)8xJs<$r)ZeNBFKsd)NH}qGAm;|@ zajWm)mD~U}>M74`+7pi7CnV#KvaC!f11(?;G6_;lD-ocyS~Z(sAOjI`8T*}l9MqJi~234 zgZTHWp}zYGveC{zAk2xD>`VE5pX+vHEkR{N+#MYGIAF)~SN7AX9COpey4H!(*j^fc zsfjAG_GwjH`KE%w-XrUJnhU7Eu@rO9O`$u@smEC=UjZ7;#lP32c&%{Yi|%*f$7Wt0 zybX_!h>T+vCFOg_@!lkAG$dGZ&>iW@Kt4^JNQRJmf0zA;c@LM9Do|Pz;P7erm8Slb zt&5!cMtO*rjhxcc^ukj7$-nG;G(zDPcgDG5_zfk*Kc}=M%~a!e)KaXjOH@XbaTE3N z2A%UOR;Eh!h(&HC>~re^hEe+&YKmyYvLlrbUe4BAgR$d4@*xVX3RpSzQs=JfOWSqs z=h(t#%2d*QCT||M<}WJLnKa8`;Rgz<6p0WeJ3q-@Q$lepH{+C6WY=cN-<$ zmZ`XNKwbN05sVG4LCm{kUfc7*1vYhHzyoOAAy5CYH@042+5ArA@RPstoPXotVADsz zwmpM^5QoB__F|u}0}aRCRdz;Xw&2`65tjymADvpmomVQ&x-GRpZ9X%9ANv6Ir8f7s zawPYvNz5P2aZf{ieCBE=gc603h`t7Un^&dvYq)=^gc&cEVQk;OR^aUzIKc`DOD_#9 zd-T@p7qI*WaDinU5wGL+vLUOwbKczLL&oE8JS;F@(lB&r`#Y+ZU=&k=-gtnP#I-8~ z0N(EyU(_-c(=f>_92RmIN3^a2BgXzM z_N1|ky^{F`4eZ?RGU5$Nvyt*J&!+mZ?H#IMorDa0wnOnZ6AxsVTM|z1VHt0Gqryw-sQlQ2E4@66g4;*eTu*v;? zxjzBJY-wmla`K4s(Mtz1Myao6oZT9#Kb@}(xCtpi5_vDe6Wd)hJ zTLt-ZwkpZleNgr0Z#Fsm!HlKNWO8>5|C!!P zp;wCn_OGqysK8XY5-OTkHax;Lf!%${gWjGuau*~q$eQ&abIqANM}d>cp~(lKgx;Yc zV;7LE-&klXo&YYo)qOwairkChdOQk8Bj1%ApN$agLh?(tHx6K9oj1tirEd}R>mM=u zlO|}Z2N|cDq-iKX!$#8yO0&7re6SpS8|RSJvx2ZL5j7~$qIC_ifmg!k8%#6DhC!NI z0cS8CR{eIb&uWqP$tl{g?^BXAM+_)%kAd$)ve%Ejao?cql{{E;R1IMLe;?}-Bn5-9 zLOXW>LB66k--ahDrUAy$DX0(0H?#dZ?&XvRhEb`HUn)hvOF#WtBZ!DdiZ?DJWt5n z1Pqj#COF@RT%JSjV8+lmX5L@e5zj~UVnNgk${Nm#`s$DJxCUapld}8yaL;sT)^J;f z!naySXhuq8MEte+T&99E^4k*a?4MbOizo8}X;LGslvHb!uKv*&fwa9G#i%AZiP%Fb?bW?r5HN!U$1Yn%DV9xvXF4#+&>TP=`Mh zOPghPdG?F^ahoAfc{}>U*9`WijAApnqnqk%o3xU9BB9+t`oFp3M@Q?gzA3u>n?$&< zLBRht9sdL#_R^!m@{|Xi39WGi3T4}4Zvh|oT_pA?QqUp2!?ru3AO~;2{U{%4Ee*hS zK{TTT4wNx4>wm?QbXU!^aobu>Wj`?Th!wD7NzFg_lXn9qm|C_6P;s1-jw8j2hEP>6&^AmqcOj6@XUdaGtKs8O6?MYt`bM5X5ERBOKD1Ksc_BRl7iZyMEH1!MJz0y|j~XVST`pYj(5dLOfc{q@a*V&4|f(+ZG*zLPEew=3pr z`SvTG%kQ6cgd)k&f=;?WQ$;>^CHR-4aZGsS3aJR~D>dGHvB5om8%rI(Ilm0rVx_VJ zW>y87zfuJ&DQ=L2XgtNWM_OBcsL}<>a+Y|d^%Ipm?al_T;3mZzf3<>XUV}9SJZ10G zKrZ}cf^?C0AXwXO*6HRb3@t3Dalg*o)?p}CKDqbIX_bDT5#*O_fzF}jg>a{yT)zIM zMPQx)r4o9U28XQU4}uXI)$zTkKoJ{3+cZ+B(Wj;e zk1h9TUzMzV{80$Z{w4KMq@fpR$oiIN%Kb=cby(8F3%ne}q`;v?oVcp%1&>+j{Gn4s zxi`Z6f$86D7@BEbGzWJPXIumrcY_iu73L037way53*LAMX`yrCpULx6BwrAzUN|7b ze{g4b(hpF<9uC~-*8DlSubMoYfF%`Sa*gH&!{$#AQ|<}&30v;Lx2Le&Ts%F$yx4Jq zP&J0@#|K&xYA~V%NidewGZ4N&5Iqtmc zC*JM26(os+e{G`!vi3<4w5xZRL%?UXBg}(954j)rnoP-~ z!wu)Hz4ME8f+yNJXgSy*($t(&KlRM3?5Gv7Owi7XSK~yIaf7vF z`3Z2I?>UBUXW-|@>Zmg$-7?g&<)5R(>W(y+?L z!)SH*0XZD_8lfC2%hZqaQ3`RogB#r3ToBG211;*JdQX=R6!YLhAam*!8$e;qWPE=j zVc56Sel^q-W%&a0X5$H9Vn98foXD}zZu3WvvR=!?*=96t7uzcv9*%+5$_6sb$1jlX zo8NPg9AAOAzZ}l}PGw$JVS1WjS;)9R(XL#|Ap_@%4AYt4O+EXB3+o|;s^_hC)f>oi z-I2PCTcCjhG+Nb3^;EW~g`S`z6&5F~d~b&n9=if;F_haWHs3R1cHrg7iH_4yl=(L) zm39$N91VMRTorUWsLUW=X{(5!RV;R)eDjenYQ`AMDY960h4iXx^2B9>gvoe0s!Ezq z%2YH3F!sjYPKjB?cKGLG0y#tf$2RI}XE`)3Q;;BG2bi}Hm891+(grCHn#K>J#UsR_ zpLF6Zo-`Qje(U#hGpc~70V_| z!*ehG%vTieNjWGhf8?I*meLX+CEFCLbf^{g#c>?KkND{OBK>(-3`~gH(<8>T?DG4y z-~(W4Ww8u~KhO<^Xh;*dZ4!5$F#ph1Nw0m5_&MgBG?W2w9R;YYHc(orjFkp;C{m9+J@|XqmP|Xwh2mg{?KiNO1`s(Z7|9**OD*$q8~TGD8|Ao>)14=Qp*o^W3HR7wmkm0tpsUqXI5 zhO8d*zco%RELd^I+Qse_u-o-m()vl5VRDV z8@z__$LAxE42O29Ezv0%liq~fILN>U=96!A4a~O(KDQfDo1obne;`6%g3(oj{sOEO z_j5b~Q<5n?rQbC1 zaM3O0a+R(hsB`z3i<2ae2Oo>%3P%<_Rucs?K8e(I5g?I{IE>U)O?72R3h#y6haYx} zhcNYLJK62jF}go)iU<89&zE8KSKB{w=dpfL8YG*%1IIi&m2h;mxlA<}s^gRKR6?(+ z`huKn)dLx1K+tcthPmqXLA&nu^x=@ZwT-!z@uSE?r94No{9cVC&_XJ7{S-vf_d8S=+?SEGJihUbvninmGJ5+Sg7+JXZN zB^;$=bNC-hWiHM|vw2>>Nx~IU#Cr(BMe+8ee=&L%MvR z`IR+)r5x!7<%&HETFehRao}nENH8dRWMS%qQZc9z{J)^}ysp>zuwt!-M2#ZHPJi%R z+^kJ`-gnQlXzRxhv`Q*O7I=z?6k+yt1&*OC+QYV_`N+n>(KBK4 zzwn?ti0tP6un;LvRKp>;N5a6fzd>dJzce=v(cXLya}~8r)vOAj62%|fgY)B`2RAR2 z0bDc80j&zwP4Vdx`cJ_c9N0U@$HUju&(jw^!Ynd<+~wlZ+~FG2Vo`Utap2)?^QHU; z!ZqQzNWuwviJ9*gGxS$?_r^BY8d8tpPpEDW(uH0mCNDnVA_<=|og&FARybxE4Q$7q zD`%kNDM7Up`6|5{2D{8o&Po@GjKz;EUT+D!iZ`IheeOnm z3~-iJZ*~c7g>Ac)!sb85jx!#K{Pcsw%Sh#JfzZzK@IGn!4&DbLmGu_TcGrv-Xnzpl zG~=}gVVHEeU>$a$0Qu;PZbBR(QL$jp57claT^P0)G=3bb%403eF*R?_9^7K??Sti2 zmB!rCr82bOeB>q`U`q|^2~ODECQt=jVWtnDxY*x93cdV|EQN~I2=A+>@Ur+fqDF=0 zB&eLT@SiC`3bf-O=g<86!CKeb!M-Dn9qf-Wdo;d??_MV(bgquotJS;`EpRPCx|qyb zHw4b!QPppJe#5E*omHk2#r77lJ7n2hA={PSssba^!CGO_ft=098h4oq11~sX@%F83 zRO<9W&()sXi6oI3bw@filzegn{Mf_`k!AKO4w5i`HPat-QX|$~^uvky*9H_0=^DiO(K1FxlFArA5 z9|q&rdz?PG^_9uDB)iS)(HHi{HT^gz&!i!o^t+fZ=tu0xaICsSDFqgQIn znS_r)kCncXL7q4#1J$*(AVW%?LHNVuueA~fWoyeyoIF0&eJcY5dqt+?Aj1pM?7=g7 z7Is&mB8#0XbQ2O^;Qu&-%c2>#;PO6W)b$r+Y)ixMKio#@i;=BBLReR2Qhiak9K-Yy~ zbh$eGXi>1*fPZi+Trs~W(-+Ar6Yr{xctNO%m-w*=4~YHh%&ZRd+NkeI@rkCjL|&DA+TS7!fdb<(zEjdjTU z?AR}sjrY_5&k)nQV10oP8$@q`J|py)QXkwj5lSRFxU3j-$(?J5=Ywr*vgO+wH$JxQ zbZn!A+*<&6Ab>}Hu^poPuZ`gTbsNt>yEZ~!-wNtJjN!O9D030VYmUsTd+j&Fh=l3B zm&utrxUFujeiR25uJR*Mx8bO(68i|LX4=2^Fa*)J>I4nxYzbtXlMd@%JSy!O9L4Nm zM1X;tBu>6vbB%%E){!|jv@XfHIvu7XZ3<$@kH<^2L~!|pbT*%Hrrf}LhI5%67&9rK zF$K9;HLhV3DBdLjgXyTp327L`o(xwDD0l-=z>Hvm(V zJz0Nz`%EdgOEK~HN#NEjWFTd9V{Y>o^-1MAhWfdMX1)N8$|m1!`r$EbU{RhI!g_-t zJc;VAGdFlIlPrq`9t$GUR2{%+nk3!5COC1?QmT;3YoCUy? z9dbWkrkkD57cu3FbSK`pMTUxyOxst=L{2LM|`$+!rz&51| z!;AdM^vJ*^((fw_ZDzKj0Agz9oAf+_JHz6m=y;~at?l;_JX>!#Pm#8 z((J`f%@Pi1!TU#Wgi!WIB=!`Kk??|vAU`fPyoXr!jn;uGB9tPyak@;6>y^#|Pwx@x zjlI3U(jnZla*E}pyU;t4!W!U^Q^?Xyv%G^(Lp?%6X$rteI&%mcmF4fsZE~q;OaW0n zH6MWdaWQs~na?*j))Toj#m6wsLFz`Qb<6 zDEkm&qS8OF7C8k!(17ZMKsp4?JIUi-mtx3rD7(iOaa}X5iOJBJ%drO+zlT<9FH|En z9ztwj*-kHjixrm)e${mdln%d;xPEk*)ma~0$LnT zv0iU3u_mn#9xtYl8rT!bWZ?A$kfE1Le&yD)n`N=(O01YvdHHz5w z_k85Er%Iosr{u|vYGrf&(KY9vuZ=dvZLkQdpYWOAqo5@?{%~TYy=I?p42Xf=%=ILncP+1LI7UkA8QJDb_ z`3A;a>?aJ>Gu8rVfmgW zBOwEy62`lH>s!e`P|@yMF*q#|mRB2@vsfVZ$F=l{Yh{ZkIeW3dM5TCa=xF-P|uV5zq-Oq5L zF1nW4Pmfn*02y}ilk6uS<9++(OluZGu<0(f#P|(4!5=DYa67p(+|J9S-MYf%q!X$g znfnX+UF!X#qTrNbF{!b5g4=FTSB9=Pi3M_R?l}HaO(qb5_mYg_f@GV&-ys?QMC(dC zLI_|e$l$=Hmzu`1$2QC1PtlY#jns2dLSeSLQ67$rkVjKF_?Wb73frZO!r#(J!;+eL zkd3h~=1%BG3ZFlDf0zf5#@{kt7t4Vm11%K7>{>+*!LUD&Z+{JEvnt>2LmZ{{`XSFM zoVkmWE$?`{(rBVe1~~kfLf-c-ww3jrq#O58YK$G^tbOFYZo zz`3V0s2b4{lIUc_MDiRpjr%xYr^LdahSL zjv`7lZ^3+%lfo-L*YeGs_SyX^KjLfs8#`vl7qXX^+gQJC_}NeCax7bCqydkZ)Vo3M zufjqGLhg#(L8fZRp`4Q%Uy~%+6H2CTb%Z9NoOu9y@mBpSC8@fQfrU{GSRJo$)`XBt zfSC(xxt6F-D$>XT?E&$svN|uM6P?qs%Po>d?W@$Ep=PZ*M$ixeJFOZ2v;;b`_Qq?b zE^Z~d+HSk_h6RKPw8NSQ(rfw(Ik$2v{N^eHvIvkW4hQgkz=g;ml31JqZW*f?QDq^r ztHW~Kt)y)B<1N(=29qh}_w<;S>|&aG*Lif!LiZp0yO1WY#JYBfGE=y;IaJA?7!{{1 zMReVr=%M7K@rG?$`EyI7kX3lbsS|^UwLwarEt*084PsrjUNR_o84E36Cg|31#kpN7+-)y2N%Zc+BX z%`)im4rM2NE@rm{!uaur?;8!s0lX#I^n?}o(J8hB{)dXd##^m5!0cI)#xc zz6XU-&njChmSIiumz8nIsIFpbIiW@VM<8Sty{jgPec)srlnO`oCw5z!JirzJQ2 z+SO!O2-w{*e%Fl(R>{tq+H}&DE&JLzmBj8lv&s%C`BHB3^#h}n0umck_nleTk0|}7 zijCzK4whYf{EyFHdnqt>%LRyuWf?1Ysm;s$|D$=&9&hr`9l7@>J1`KV7EWi)33(l)>%0of?GJ{kL8|s9UgZ_3;UR>{$^`~+0FnH6qMt6*pMGZa zeE~9|jDB{!;;CRd zeDC7hfKO}6%)(DiOPp>%EqB;KVp;Ss1AkH5Py#n}EIK6`FPgE5djR=_Qek1@G34zl zjgOfVn>%3g!cItSGvAZ=t2H24&(`#MLo$vbD(^}bxJ;3Y-Tn}Rm(3$sZ`Zf2Z_sou z4!QqjZ$A(X^GPw_L`!5mm1AVA3yNtB0gfq`o7i&!PQrZM5gwN6B!J>NuslIPU~xEX z_aM)MKi+iW=XEn&Jr$RyNtsT`KpTmInx`4TO@O0>KEz&>!-<^PUqGQhD>_Ne+${-} z8BIcwyDwalL5tYns+$GIydgt8PBqNUyCZYGGZx_8H6hL_pY+@VJ(OM5jY3H*RliYU z`Zh$k;`nu&hUKHf!Zbxs2MgKeTk!X7t(5c+#fwlEddf#5jSzBh0h-iyhUH|$IUX`j z!WBcgcGz;yqJOWP7N{P}q1-=amjm%dhUgG!+6?VM(4=C~ zqBOfXrTX!zDnr#l^naz7aUujV%?Zz58PjSGEYRPRR((cZov9Sol#+BF+NIEI!G=-g zx+0s{YKKL&c|Wp-4+c#mp8XLE6Ky+?6LLIfgDN5b=P&0plmj=zy~M@2gbhEL*QE=I zT%VSJmLOv-Km7nBtPWgzLV|UK-+i;RwmgNPoMpw4j2{WaMsp}(y)Ml6?Q}a)0+)h> z#YzT$6TsFc>LztOwmiKaZ)8xE%SORTIQ{QsQBW-zYDB;`%{ScJOnCK$cZDfoM-*`` zu(={RyYsZc^+~jrcv`Qgj-H7fF(TX}Lt16hmj$Oq7n*|rv6lHi2d;j__^0yo_{+dZ ziTf-=s72N-Goh5)Bas+MR!DFTfLXR>ew6IuM%+Zo0>s!S{$h+B>n*_yb+ftmU4I4W zMd2XbH;{ERUd;*V1yQi$UxLMu72Wfdr*8a#JwxyO12=X02OF%fd~uq2<7VnE#^|0= zdHh>pvM>pjeCm;(5ZgIYHZ;lWb6C{Me4Gx146mp==58-{W`V5B1Z&|_@e+NWfpe8= zKy*6s1t5u9>ltARnCZ_IyF869eWs@aC%JL3L5S6jol`OVnS*(^$=E&PG(D8fPSl{F z6sw5a1<@f5{d-&BtF(az^;da(_atC~r55P=G`h2ve2#}trh5;+*K54owPo1_+RR+a6QHQEZ`7fX z>yJhHK`NtNZ3kUNv685~SZNzQAk3Ka3!1N{`?je`s?`_6mWikZw@mM1q5dbputk9S z+^au}kcfOQKvP}le!d2WKQ^>tHqOx z?~u5#8CFye#}vs+6efC@E!!zWzty3ILh5{pR-}MEXuaV+zahmby9aJpz8&R+oE4Oi zy&mRGoej(^9uQ>7g3ISejkeV)SpxfX33YU~Z7`(X!E4&o3J!)Sb(b@W7c02B!0zD4 zic=pJ2&0y?yagxW=&~9Y;%ycfVREnRlJDPC#x8rFl~K!t2ns&dVk`v|es~^>ZMMdB z^F9T%JV+mL-zf%iWnN5!Us3YJAf{&~%wGHu+A)&#m1!L=s9i8Dp>f3>sh(d--rmS; z$iA0a$-SeVUn}O6Zz5DykZmwt40i87FAUMx&c|nj**_az*>@9v`6h`! zsn?bh6$BUl_p(?cRZn7O(*;D|H}LN=5IiB%`coc0EvD61-JI=n+*^s6?}(U&cKZw^ zuuSvV0p6}#sn;Pn`M^S_J(U;~m4!^Ke7Kal^4&qe&5fXc{ye1xe+kq-s}fs%_js5`!gn%9jF5K-jTW$uU4&wP}=DSOZNO2j;vodm0G8z$&r z$#yy!hoBpv=i^PP#e|ufD@SHHU)4e(V$NUM$6q1FKPta6#`*{Ol$nO~tDn}9HMrM7 zgHPmFWD1cYSD`--HuYgIniXrMM3s;+`gJWzeA{ zs)seZak}H^p-OBY4r&=03FS09D&A z?_QiC8mG*x@5nv{Ju9$QSxN_w?LdlvMQ;3CMJY1$zZxvlb`#|=gA;WpoQHx-gIvax z3-0@uNi72_F|>FMa)3sD$Uj)n+-(rQT|ZnfrGL)PVy$Zz*n1I6MXW%nvjuIvf|5q7 z(U{fZXb%g3As*l4;;G^MgmATU83Jb|*E!S|;p?{5!M38pB26f=N-`hPoxJI2|mzpE!ckJ*Sp zXaM+R{8Bk4&jq_pQI$52Yd7cg)PS?*7r$4;^Y-mL%RYe->5>wwN##`{hw6frKiP|= z*uXc|r%7}!Cos3gsReHl-Eu zU8icTdq5N$0^x?Q6W^%&W-o9qRS*LZEBA&(rhQQ_{d!8xsz!z@S19wq3I5GWlE`C* zlR2XH4w_hoa2%fd<6v^xfo6vNLVH*=oz#>|a&S*I@--wznj;GLhzU9YXac8i_}Y+Q z0b`uoR%k0+02oj8Fuz}ZY&~qKy_=W{vm>}@ml%iliN9(r>}?#r==qdN&M;S!C?V5J1rYow9XLM(I;@shSf`K2i|6 z`9u3%MhGor(omti;8$U)o#0M>0!R%eMR`EYKXEt}V@JCvqK&Iy7-B*#DHC`)jpQ@EQX6Q96hulxSf&r1C z3xCOQ`GD8!pcY0vvt4!nLUdE>(OWCf8*Z?iP2LN+l5&C460zaj^ci}m_nPMBEA%j{FF~xk@a=fO}5At&}k9v3p_7*ao)|omyYz>VCA5D zV-byHYELwewD>~8vmtMB4@VfEhmqNd%N>C@3<2t_!m@_Kuz?@4My= zVlb!zx!ujsU?DT`L6xx=Oh~>nNtn%)Ris+a6MxaTQb6r5{hIvS{qUoV-l95wCQ&#L zSzuoizyQ3*D=cUIjQ2_g(MBK%%0u6jAMG(f0NcB3bTU$jdVVs%#lOdLW_CVVu*jom zoSQmr@6_&aA~js(&;nR0Hp-j~@rw_&Ikw+#5Z?BX1~@yDARDr@^`FU+uQe5{3B_e& zU44&c7k4t^Vh3`d)I~kX>pm`6{^NKTVoG=L*EK}7=Z-OecMSL5`H3LwV`DobBCL5j zzmCFkaV#L(Ir3c`^2Gs^1?JfCeC9S~1 zpgJ;X?E5m&WTafc0lb;WyXTzbaNqfe^Si5mohBsPZ&xE)#mz4`7)J@a&)@HFDCu;b zbHmr63<>yH`zgp zw0D}kS{1OLA_3VS*tNpp-=<;!&TY13p11pm`c54m z5iyUIpE5KrJ3B0s6fp8KGf+4d-e1f?1eb3#eSdv;bx~L!ao@F?Fm4pkMnFg<$UH40 zlxngH3oeM|h0b=nZ1ZHK%H3`4FUc}5k2+4-BbY066lQ4pZY6Ufcw#vroORoYCa5dx zM4wI3XP5H6l2dJXVIQ5pJ_GqL@2Dg+KvjeXqVa@F4}(S(v73`$wj2PEQ~> zW@&r8YWx?6RsKFYT!2Z`tmNVV)A5Tm_$Vz zgZY?1E!7XzHukO_iH-qRy&u@DqWqTsWhsESm9pZh#d#9F-x!DorH06!e3=opH-m5z zl--QZHxTK*6g(TX50nuI+}%l_;JJO9`3r1Dqc4wp&wVZ6o}xma82smW$}eK{^;tcH zIP^;N?Y%M{Ii|-gtZ-$J1b&kF`J` z#n?;Vmo8=j8)d#@fhMp5ZJNR5?F(yNQ;Sx?OE}=!>?v8E-C=OZVK#5B?>r!L zyJWZXn})f$S)mj>$Vf`8%RK*@O^TK8Z(`|RK$dswEWZa1tSlGIT{0Wo~nO+)gGyyaZJ^Nn{P0dQKo3KsRuIYu^>iLt?c=(_%X8A>GZseCi-b@iDa)O0N7Jp|*Si@@%0 zIjju%I3JU=Q@paTYE%J9IR5T@8Q6^1GwMSqdzTw6U}Ug5BQ$IRgQ>VQ7Sml?*U%{P;gt zw>AV|hsC0ivM4$V#K_4Wi|4;d2E`EzyD%d;bA;BR~GbN8n_?7VeMl0yYMM8UIpvPRwW6`-B~Gx= zbdkjm*U50xXQl|cUiD-WsWDH=i2Ju&bQ-ybj>*=sxd4K~Pc=$sriPJ_$IJ5r;pi_a z1$id=zE`NRPd!L(ghXwL>0ukm87cBkSfI0ozxb6@npz3-A&Z=p{x&HlG1eVQq8_jO zJL{Tc?%N&m@?ztLb|;ys>KG?7u8K#4Ngm0VNAe6!WN7KS>>)3~07iiXu`xF|Zft}T z8UVXP?}wGaDI>~fz(l+1IXI_Rundv0;oa;E0v=@D6aj3ginN+YkZ_vh#=$!cp0Fl{ z2i$h?qN&_#q+oVXGx-jOq*b6AU=?C@y^g2v&y?e&N#3gS=0|wF^bCvA>{X%%17eL= z`}~RuxW#&NB}R@NHF!#AT;6t`BR(FP)++C0sREv^zGUzXv7@F>jQt?70$BrIexV_} zxnkKfu6u=SbX`-G_~_^c8%wu(&(i`Cb(txzT%cKs!bNU|@6guG^%ou!JSG}h1n-RE zOt`ERntkR($4t3l4MxIDkZUCa)C#8D&7`-irn_OcrrmeW@1Z_FKB!Our5ScnH#=cU zXP^2ITc*2HhXPK8Kkm0L51^;Z_~Be{?;C!8rX$)t>)aLbUXeH9M42bRdzgjY$4h2 zok?3WMz}eoN#o%NsO$h4Ke^36pGl8{RB}okDUiP#31sRXMst>k1c;mho;U-rGSU|i z5Q^7S4iCcG4_xl8_vcVxRh4FBrHM>1d=6BHYr@w9Wvet!Ecl%)5L1UhO#fJdCF7Z6 zQKqIgg-=8r}DaOU1G&#Q~; z!t2LXw6|L#UaUNj7BX)4=bGL3MZLe|=dh4XQ!}(hm_SDTPKrKN_d&#CprcfhJoktF z3>7|(w+x>mss>7F1T70xMlq=^vx>BR7NKORmyK|J67Y(2j?4+?im-(E({?|MGZ|0o z=d=ROpb}4q+zMXeN=edzG;^!qib=~ywVMjJFvA+^4R&SHQ~M#kf@SY^4<@Il+l7f! zYz>Gr7OgcU_HMifHXNQ!bP=Z*A&EaqJ=V_`Da+yJy7yhAy*+TWm*G8Ekivf3DO&%O zm&xr`LI&}35> zXaKO%>d%*VFx$@i(;DBWkT%2w{`vS&HC=K6#2^bxFujZw3&?k2c$)%riYa=;s~3FMnu3ycyC@ZYyCflC#lVaqU>P z`Md_vn_e-;OfD*&{w&{TTB@Nzb?D5JCXccTxHWh`WbyRI^L>;8TWh-jrUr=dihCVR zvwL1%jM1Fd0>~=%1M0#e&oZ#@#h-+Dv5MjbtQR)o;muJXzC5~5m?@?^Gxg?%z9h>6 zXs+n1=&P1->7+%67Te>)L#Z4*lN8E|zpXD-q}iy%Ck*6Zr!j@g%B21SHbG$-YhX($ z#eevO7v$0LufTSu1|if1D2B60NgnqEQ!#RHeCCCIVYlS=Tv-b#?ydVgg2&*}zaB2= zxht3;mVkX?hm*7JewkIi@1+%)^s_K$%pTPXD1wM6!vv)rMxgxtfQI9CEK_*O`9~UE zNconK_)wX^a$Yx(Ka_nXG4Gf8K8z3QaE08_d16Xx_1{^Ta}VgoKq>NQmY}0@5062) zKalQNu4J_79LO&Slv7)|HZjkjGQug(>@@*$ye#cSN?vy|HbwgNAma*?Jq$h+a+mL4 zUF?E-FPk+hg3gc`o~LL=GY@mCbmpSFRtSWMb6>8RC@`%jNxW32A@+RWpm-VRxcI{b zA%j0V!TOEUvL+mIG8lo4Wn(k^$n13b5)?B8Kd>7pXhmc2mvH8>7q`yD>U8d7lz{RZ z8wxd4d0Ox3+04OLfmM!5&h##yhW-&k4iwm&mM*M(yK6P24A)iK8$u}}kp_wwf7DWr zwXH|!+=S*;AoQi+{wK++=QUkxlm4Z9``5eoLjg;7J0J4v@d&J3xx_v+-)|?-wA;;KI8RA(G^$ z#dX1=ofF~*;3*~{Z%{xbFz9RMpf}3b#Tk#cty_>5MYPok>~lhjfKUs8T?2=N?iBTY znqhDBII5D|78&)SydGsU8ONC+tb}r_BvlJ(vF!-1k@O>8_UJrv>QKxJ$rzNvP{fVf zKlkoVkn|lw)Wq%EY~S2Rf-}uQRZ>DkB=mcxZ_@7t*tHMb>}+v6av5x6mqgfB@qL?L zeNi@nppZPc?Bm4+`IU%LCI7U(89d)syq;=Da+Mz8Cj;KJ$3@V@NGWd|dG>PdAr`Ef z9*IDtiT@qAMT}3$n0#DIVOlVo{E0uSRMyzb=D z_;|qa*&U_fM*T7v>YRYr83!Li0F~FH-cPSgEDQ0(XqB6D%S~|g_WjUvz~v9@-Fx#I z5;bN1%T4kq(5r5pqkW6(n5q!1=YdaU5#y3hmK&+iNA)FKvN{vM@DBDMd% zrk|-#A_!B%B}m^pu-$g*WbEE#xKm=NHQP^YUZ(vCdYqLv?D?p%)=6UKZ0_J~Ig5hXQTK?tc^d?uc?L z5JYe;RLeTS9Y$n3uz(VonOH5!g(4QT=v7Swh=|Z=gY1}?{FtRr0&$_w&jPeT8@ER0 zl$;lvdiG8a+5vIo;qZJM_;;HowlbYS;;~0*`{C)2uLJWcC-w0(GaJ8>&anclq_o{j0g%P!(kn9-BJ(u1QKYoBt8uSQhfq8z88cnNw z3)il!^zw-iR=e_?nSloGrVPSa+#G!QhoSfpAS%jd@A5j;K-O40>{=Ogp^#f5f{sTf**d2~I()r3R!4cqsD+kMGO}09<;R|IvYAK@_ZJE~}kNyq} zb^Gro*Ps&Nqik2S^%nT7;|su>!`i+$(l&g+iS?dSDo};4GjA!;wFmR=_WU_dJ_1@z5^lZ7v+|TFKCj~ z_agw{6m zQfmB$bij_?LfQtjRB3Efkgr698ON=S@#DsHx?V0o&(8k+h+3@6hjh| zXvQSb04b>X73{A5Gv=oRXbEB9vk%mg4>XzXX8rc%Xw^8=+a$wKyQN@I1IB)BU)KMGIYpS7bBXej)c&n#u+y zIjqBFdVi0tXK&IgtL%5zWY-yG{pD2yrdSalH%g!D^mAsTsD6!FUe&VmMiGlBs4E9n z8s?SwWJ*J7aGD;~y>*~qw}SCo&4bU2VW?Z`=1-qf1OAh?H`DySc+%9yg$p0Oj$cOI zIAT0w8$N)}vjEt$0(d%A#t7JqcP;^5CGn#4RNap;^H)m8vfQlM_KllH=$fo~-eT@g z3A}z#hjNC{dz&9jN}i|Nge4e++Dab1MLA@W%Yp+Z*9x9^&N-b1R96n_c0+3Mnfuv+ zZ%`DnwfAf28Tz~DSsxFZw~nI#7^qbY6fTcc<$=W#mrP8;{&02$n)(rbfD=&^u&vy< z=2c3W9*eeAYuo5iFe$Op#&!ehTHo`L)(L9o2bm;~aLGe1?29CTH1?(>6C zTt>oSlPV)3DoJPf3B?!9zkh)Em1qagS z4n~tlvpH)lu^;!V4n9~iu0(@n3?4%CnGgae;FllL1+_eR4Po!DnVB*tVs!u>cl^UU zVJsiXGaq8Li=JJU4hhlfSE7yLz1)3E02Ma=SZ3k%n<=Jb!^_#Ky6gDHP^~|G1&6?4 zEE+H|?)7P}{e=da+#yz#ZYf}5fB9L36bvZ#M;-2(&tLYM$VNxFQNXV00&WJe`g$#iv!!0jh5-74rCr<%{Cu-rLoa3@y(E4q`#C@&tdhikU=A z^(wj}n3slr{D~QuY2^W2);Vbno6PmC^leo0&8~-R_y^1DHv4%eQr|P@&j;nRlSA*Bv^uSOuECv^s~mHhT!--Q8h6 zNrU9I0nkTh#N}}^2m~T|+;MBg^v0`m%y%z=Lue47?ZsakIvzb1@S1o3F5@b?fnzp* zf5t0FideU0lZj0E=>rGqbqQH`Ny=Yuv)(t z$Wh!Zkf;>E&X_?fZw`5cn@`_^QBLn+ zKk7^eH7Y}9Ww0&b$7CSXowl}`^nIMm1bKK8z+Ov11yRF0LZaJOA`LplG@$aw@|#`s z@K3Tgb9yRkpLT9nGFG=m5JBhsI^x*b-ig~*J%>Qpe7;##{q&xZ zi9?Vlgv47`yFIQWJWBJX(>h?Nd&}{)%iUwOy}Z;POrJ{c=K;OUx{Snw{64J1Bm2BN zW46i^Sec8pS#?)fp{sGEITrhSw;wJu*P48vNr^dbszRhcch&=_uuT}#pTgKSgRcc9 z%33SOi&gp+6BAp*hTx7>rIp|_6=3e3gJkLu*%%>`gXDg!YB}xY`lj)G5&+wHV(?MB zk6GuhF;(dB5Vd;&7>o^=rc`mvK(ZDuNnIIzNxkDj{17}=sl zPyrzZQSAI(p2V#$fe6@XiXBOVRY%v0y|LE=1aSoSBM?t&4&n>cD1j6U855uebpuh* z({dpVmV3~d?NUG6kEny}Sc8Msh>iy?Ywu?=ysqH5ST)G@oUh`>ngOgE5i|F72}G5f zbT4LETuPz(_Es_x?(qM9>QG&7#sZ&0{IpKN&PrNZL~_U}Z_qBT?xWi(AsPG)G}>IU zOqNe?pR9EUm5cz7!%i;|@0dN(7Zrh(%Pu|&kK>QMa-W;VeoYV2W_LM52nIT4+fKc0 z-q7cxq&V2I%NsP^=TB}V`!9aX>*OJ!s?b>2ZJrovcT#zQ8&j|eL@yt69&*@VVeh#g zkMK9c85O=~xVcQNUk?4e@eWps)uFQQI``c)>cw9cQ9XCL2qkFn>h#JU<&4}NaPqj1 zAp=jI&9X6s&dMvUo38K6CFeQjlun{{x+qH+K3Wz>luy3>p2V~zCY5H*z4?j}%#hQ2 zGXYx}69AZ82GYLGQFG3jkgkYV9?@_d-)nv~{-TD+AG2S`L`!=waM9vB?Ya)A-1!J9 zs-BG!Y*Gz34v-cqz4Hf41qb^``lmRNFxUy?Q}q;bRsgBJffdL0o?q@IE0|U(JXik( zfB<&$Ct$AEYjiu^@rm6mxqO+UYY4?HkA)2I8c0JVrTWL_VF!Bkl8A>r~~JKoqBE9 zM|m;@N_~q+xRDcR2gv&n6F+f`e*JD}js6X@%(bO1&Y!uRE)$xnCzWqqqiu9JqL|C^ z3qf~Z8?Xh_N0qYHCeptslzhZBBxq403PQNvwKP+09}F~-p<%(tH-9h0MH0D8W-j0m zCj%yB1%BeTpzJ2c#4D;1w=8dfk^#-*j0dxSY(T$kHbCY)Sst-vX8rW7v1zG`{p)e>Bs-hQmzd`?+mM-=w8ynGXBGVVN- zJ`+PjG(3(D8U4$q3n3MNYLLsB@v!0f!*%IpsWu!J#ZWdG*ePOIupE%d@3{%X;F9rW zF-K|LorU4J#$6XR(1B&T9Q0MspOS5)*S#L3#zKXF?0UyMiP$yZSWtnYP`zgN#F6QC z;0wCF?}bqd;&x9A6}K5$I$;7+NH^!!fLbj`;+Bl~Z`Y_$4HOOFEhHj4{lNKK;q^%_~r>oz8$1fh_ehLsJEai;RzwN9YyR0s-K`3GUp!bRGZeJe7cAG7ca67NbZrd zQi@Cdl(-dzxPEQ?Y6l&7;M>9cEFz)_8uHw~$hcsZNqeYyRUcm)@Y;y-@k-Ik1s-;)lUm{0G2ZtiwI^o)gP6 zttSb}ns#Gco<6r_ffBp!p!OAKY^xFhYZ2UVQiOkMw{l7 zgY}%|Q%_Ifrpmx|*NzzEqBd{xuh_{jzhj%>bxUm zzlwawQB|&0YkJR37ic+L(zq3II=OudLkvZM3gsfD+k;3Aad^5AF8U|upp-P0CQ&ct znnBprMbI}~ccNqe8t2zLm@MM7Vpyza&;Vovv7BU<(ZqilsT;sW<`T=1q@CT4!16Nr zP{|cI-p#(oc5)S9VY!<&e>t(x=u@T+@cgJwU7ZC29&L?)OvVu*3Fp4Hm!p7+*=i7A z>;9Z>oHt^knQ-XI`78EF?W&`1AR-ry^HQ7=-45B6({hHi5>&NB+kYHe7XAeGH_==8 z?7)(PZIX7AZjb~?e^CNDpUi*u z`)y(m|3aa^!@B*SU+k`!)ubwtzJnn(y{9`{%g@1i0(a7Xizax+je}4A+75Vs& zpZz}l?RQ-d)hE<;{b_iXoCAq*`>$&UF!f!%sxtT<{;~D{W7p}bp^g4(h<~?>5mF-Z z@W_dA1NiXY?P34ae>n67a?pS7J%8KQCf}fkl-fU?jG_L$4)q^*rGM`Eq^p}%hN1NL ze!ihL|H=Xb@%R7!>d&O*acu(sucRQ5T;tzP{imh=I?R8qJO5|bK>Sj2r2q3t`=EUR zl>X=1|2ro|&nBu-f4_^G913QT`A^2zKVSNv_y1la(tkb2zit4Mdw*`|KOgD;t>Y7F zkN-|1;Bx-7!+)jA{AGjwZW{mN-~9h?vHt3a|7@WCF}nBv{+R!Nny7#5bE&Et9hz9N zJn8$z&Q>z6f}yXl>7Q)ezi-w4t40lz0W8zs%o;h{@56uDsQr1W|JB*<|K1Y)&zKmF z{rfEWw*fBw4_FmWl>W5mKTazCH7x&T%KqC{<$rK&+=L2@(ZcGO{`PQFhf6@;|JPsU zX8aFk^X?sh6Or9#ML;GjnfOLo6nrm1n0(Eu%lfP;@qJkJeN|wGFZn_aqR;C$igwkv zYVkefbA#av&*tw(*rGg?{IVxwN)9?;*NNi;GJE12$iEYA|F)G(^nYE#J`VceU)b%R zj|d6Y3F4K+)A=b!ZnoQ-{du0O*H+%%;0)gIcZ*2~xNgvOY+tJUic}f7Qq;?ycDL7g zFW>b2w!6U9v=5Vb*Q_el-c*1cUZ>rrJnoL&ah{^{KB_PK?X2$sr&uD+1opw)Aw;_; z0q5V*Zyy|S!jWG3-g+%ZYkH$fd-=cnpB#PKZmr#R|JoYc{nkEix9@|zoe#wE7jIh) zUvF=B+%LA?4#Wli>%YTpJKYrA@o{&n$bqnTZ{f?=VZWV@hwXG%{xlOHkT(EVw>L|+ zj_>RJpdM)DEiWL(*`MW4a5x^R^E_FuO{H6eYfj7S^#M#gdZ1y^MsZpUDe#+{z^F4VG5X+uL zbI$9tSJZq~H|=@tFCHqeP>+JgG+%vn7W326UmRLTF`$$~_#QIoSPVE{Zb&OWX`wMp zc&tKr9kN-H@aV6d&^FVrjs9w@|G1s7_v^63u6ubK{b{G$35>*O)%kFcZSdl~PxAeS zoPu@DuGQhF9L`o1YVyuT@xZusAK-7=y&oSP)x{zxn$T7STCP6lcSZ*Ri1E=6*8Qp3 zhwu4LTOeISSrQcAkrgeT%)1Usor?V@n`mGDRT+o(Bha#AasJfC{re33;#34xV|?l$ zD4TQ9bge=4VZaWk8k#yl-mp|#uH~8P%{!!yA$yA_`Wouh%bAKu3wq_w>s+POg#d)1CALdro zQ{@I9?+Rr4U{mj^4!$B;^M19jhN#oj9xded96?~^eWmU~=zMRV z4L+V1^*-!fNxxue8DYiTqT2Q#E}zcLXC@r)fT$vO@uG$^p&>2a-?JQVG||qpD(yZm zHGa=Lco{c=;(#QXt|n;h_+-)CWX~>nHIlMHI?o=3R!uk)ER+X^h{-E2ssKHw1BsDG z8v>TQ?wfdcNDPnI)< zeslPJ?QEo|2=3i{^7&_D22y@%kdw#lsgm_}%Yc0}=H|Ven+0t(HR|UbyJnOP>zhu* zLC!S|Dj9bD%@fN*bY*mhO4r?;c4}c2Ed;7{WIPoh*)^6mEz7UTb5981aqs4bE~251 zLyXZ@HjfSs=-^(9fiDFs;^I71!#fIe+zaFK(blNwg5t!*ITe{V?wicCDQ+L;+p{9g z-dH8Im^|v6L%Fb2%Q$ z0bAc?eXmT$1UTJ%*bSL2w!3!s`|90Lk5gZV0Y^FwVq8DH9IFKc;%TgEu}5v`{`-VF z%un@7+=elv7Ix`Iv%BnlXDCA_ap>)qYrGPAE6knpho#gw8u%hGwx2e4n)gZq^_AT? zx4)DryYrx-AUxnHo=Yjf~nI4 zrg&)Lb!t`y%<6*>YC33uSZvQ@iFBR5`c0{c=d5^UR~)tDr*}G-IS3nYPqIrwI{r}5 zwqlRYV@LY=8~CC)#bF7JDnKHSmy>a(v8=li{^g5yW~rbopg=Rdef4;p-H=iRaU*`iIa zRo>4V6IW)xlWR{q6XrWL;J(LT$LwlqvFJAZ0+vCa9qX&!n=1(%=kRa_UGLjsIXez? zBBH$>oPITxG_+){GJWLMyZsBGLkGkYYvd-O9ouB?B7N{kBvGe%V1 zBlvy&9!$l-Tiy;+*V+k?zs(ssKDSFW7#u>zxfp~oYRXOh*nXXJd${y|mq^}I^1ok$ zc6y&u9?P|9Cuim9NAIa(sYh%0B-30+59{;w^}I`X)hDxc-f|c#pvf(oPnbEUi;zsZ za|1mP=LhCZ3Xn>Jq>BF#5`|u-?JC5xTYy_>d)z?W38F&JayRGgBztA(fz%ZZ!+n@T zB&0zLNiL{8=_m}oy7cb7oy?>+P=U)|aV2Pcxd^x)746i|TnKsJ9z20426ij?l$p`Z zfv}qsOSa|=ZAEj?oqpF9kwuXW&w+O;e<2(^33O|*yUzXWHyUCN%DJKE@N&8gxxWod zC5ekt7y7+GEn}=Aa|qQ!$Tl!o zO2bLx8{PS)It1So;CI>?Lny2+1sgR#R6{bCgFz<2{+dyl4ffm*X3~dSc~Mv&cXb;k zSN~u){xvagkvh64M!n@ipZ<>it7Fku^UY15jPT<)FE6R%eG2a^?^bpAfO(tug9>H} zg^}DzlXb5>+!(!Enwua+`KxFYRyUiy3X~eR%sj|WaWSSnKC2CuwJtsgMFFjf_I0(0 zu`%|?O<10S=ahm2DWHx*APm)8H;s$2*OnJxK(*`6&)~G{m$~Ye zS{rCJl}Aup@H;?dcke1Tpk_eEWde)9Iv^^Icp1Y|H&c@{+oY};8XG5tz0 ztLX-cd_Sl@F#_sDnnK+6>U!OGTRY+~SA6$}xvY=iR@y}*8?|w)e)nyW(xh^cGf#?T z^I$7YALMQ!!QC{V`F_ZnlV@EfjVL*;%ICXe+cq>cR|mTS&ap)ndj08!Zr6L4JT`At zE@^Nrw@xExH~Oqw0WoYNbU9i;9Ilea4t1l#54^UVaDb=ud{nL)y}(@q5)u1@&l}Av zd`R`moc_$B*s)&2v3)wmMvf#DzC6ENBT5su9vQZ~(#{8zg0g4Y4GQG%@A=iTPvgm) zcv@E?`*?F7FB*lQ?n1qiO=+KX8=3Cyz4q*34X(+Pv8dv0roO~{4o(lv*ks88#o8|$ zA<yD$Lw3QCkXlkRbn;xFT$-&%pX z;mlqm+Dq=`Xd6tBJsa*gcL}#EK6{8+72hWC>UYxL^QU#%(x%(AIKtzMTJEISpz08u z`?Qh0q@b#gCU%WbnBJ;)pSqrI>5ZzS2EIx$L@^I(!`lv*HA7UBCQm1*tx2q}??*hk z<;_f+GHKko4d>>TAJQAWQa~GGzpTkY+xw>5wtg6ZMq!iN;9vRJ1@8oX>|3TLt%G@# zGnHF!4q@d%+VddUvfv6(TV!gH&G;xlFV8#g{xN>;lyS3G&()TI3gcZAzg>tYpM2x> zzQ%L#t<~Kjfns4mJ{zMO}nSa)2AmrfCLN%Gjo)J^>G5z9}RBd$)}r- zzU*!ztN8+aI%PruK*I%}uk(!s1tM|*0Lh$GoKB~jx3@W2LvHEX3HDJ`K&g_J6oZ@+ zBT1oJJy166d8pl2bT}Y;WM8YrrJlo~!jOG@oYwafMYx@ywxv1TeZ@uB$ikZ)I|yG) z%6^mHTxJ1&Mmc^quKaR2@9DNT>HAYK`hgNZ-cG7rsXqmm5MpD#IXa`Udb(6e{Kw;g zppc=wcllU#adQvU*G>B^x=?QbC~+>9@1YDFwmBi-HR=Mr6*toAQNhy+t`c;yEw;%h z+)IDI5R(QG-_NO6v#`rYM;$#WSWr&IzsO*gdvMj|#+}M<#fZ-7di z{gdzFu=CVG#UNCv+Dxi>br$9RG&gRqs{JSWhDektYat!3>vyQ@O{hI64&7(EsyT{o zA8jRlP$5G{Q%yhnPv$7W09j*E%;vfA9_4o1q;|g@Vmk^IFobYe!1&-jtKz!0pYDx% zL(jPNi_HirHZu+vL9^`2ZrXxZ-H;EDAN#y;kL|O$GtM5KMg6>Qr>*Q-eXXbP!00Rj zlV^XI6wqP_S($yd^`xXVAc)p+LNfYtvGE(p(og71&-24?t=qM{pVUut9UYt2S1-%RBB296Cz|K^0)FF;wl)i( zGF|IMeAM$->|g-c%0m5NYH)LQ7w!*yaSc_(23(pYi;;;t#WM7^Wlw9GvO%munZPHN zzDN5`DtnavuyEN;L=!6@loq}F`L*3BiC+|tK|^K2NIw&g3s=hxs?j=#w&hU|TfbF! z;OG2P0-Apz>pIDW$YnB?N5BJ_)VfulcHx@K8npdIpw{)wx>x?|n95sC?K1Iw%7^(u zMp}gQCo&2d=NuX!SFM21*+WN*`{dciw_;v8iyU0O&*a>=e# zWsk{Yd<&Z^B(VDlD>G$OgGOS-lfE6?Q`cL=(G$SbZt}oaL!M?;4UVuCmStziaO%C{ zhya-r4fAF|$Wrt(MUmpQmh_j>2Q}a>&`fgYI&$wQwYWzePG?Bz$#%-qk`l)pH~zNn z_o{uO#2cTbo30m4>xfeDsWr4LWIU-#W`m&W`9^9T0RW-a|sF(n(ey|J3IcFTqYtYv-*mc;?x*6~lf(D=@d#h;$ZrF$&W-6ShDaB`i2 z;axBg#>aboJw${`;DhosI!bSA?#&_|4v2!>=s+kE?)rSZWop~9A!>u~YW@6>o(VzD zhA~FEUH9;pl>O;2AF|C+$m)gF?lr%#CWx3^n)feQC9K8Qk6zK-&1b2jL9Uhe(k|G4 z6;5*85rD6giMU7rpl+3=L^jX!tkJ!G*WMj^g! zZ(>|RJ()Kb>Uxc;@{4(#jd1x4Ue+e*y+?6Sc+zd#ep@|XQca=FH&hk%v3Yt1TCNY! z9gqS*UZ+ox!iOf*MoM6v!rSRET2`({;U>DNLO>P-(`A+%tQvwX3I?W#`f!d5{0}aH z+Q72rHh`0(m+w+Hd_K@3#l|kPutDN(LB>qe{%p-ZLcZsHN5T=0TF9!)i@Q@LG&Od!KS<9W z>RiQl7IvKn-eOR?;P(rZaiHk>POgvk2TwM@YB=QF_a4!lkFviB>yhJZ((Q$TJQ|cE z%O)YK$*ZJ{H_NQoD@sWoTYAE_cEyx&^m_HfkZVxyd(!t0alqNO@DaKo!&d07u+0?j z_UZNeEGM883TE1-ssj>3m@dI>#^9Cm3?m2E19UJ zhY@;js!6Zy{B#dwJbg{~8p{wg`|fD->YNts_GyxHg($X<%5y90&&IWjpi+Bj)A#Dq z-7)N#1IvVJdfjoel-fhNLmV*CiVdAWDDb{s$`EVFN>7AkG=$jPDhc z;Yot26osPV=ncO>#71oHjf1YWp&OOk@H8quD!)OMywDy+3-oh0)MtHqa8$zuY9^JZNk zpyp!T4nY@<%bMWuuU zMvxcX>BPw^bu6w*5ft32K?&8;R8!__>Nd1EcMi+ zhx~QtZlJS47r9=hLqU(1sPkNw>2-ke34>*vQ{TJp^ez&}5OBqOIUk@J1ktV5qFJt@q6tGZCg^0nl7`LEz z5PGZc?gdNYir+L!!wWMiq|}VBSVr~NP`zBxj^~o1d=DvH|lcnjtpjg8N-=72gcpe2l=$Z&8Ewzl^YhX zeon8Cxxxi=YM&Z_LB57iYxc>qgK%npD`9EYJp!2bX*NZ=1T#3P)Jf9@FUCjn24M1#IR+b4?D~rs$66_5$qY zv)_SE&^<6L6o35DZFb_hvk~POsg~G63ItWQ!*k8^1ctK;{hG0BUV9-~kiY3Y_a4_A zerm7Z-2}Tw>GaZy;bw=g=cA+>E@YFPa_dZ?f7o?yTCIjSX?2XkDALI)LLJ^cNwB1^ z!*dH!@O4KcNU0u84ODYyen-)YyX%KVAuG~`)7vR+($ zlE4jDT^;q>F*=h3I9bKvF0fVs-ylj!5+_hWI{aE(Nz3yA2-OYSd!zqxu}>E>8Msv}4{ij{4eASH zdo70V)jryRqZ_fkcg`Jk!}*L^OBqI_>`FxiSt zEx(MfO=eq4i~}lisQu0M!HL_|k(R%3&SO$gl>GcZA7P3lI($INn zm-sfdC3se)2TALey!&O(8A5?#*&$p?$u`A%oJbmx_mYC(Xv~%Am_=`0QfY-3=~9V;jWF2tG!B-e_M!{twHzQy|1}e1jxy#+}E7}JiYH5+;nV?tK7zOG}dW1gEV26)#V$8vxi6#dMpHu!=hB{!nh z;_RNSCs6-!x^aD?E=8h9f2a~VzI-3k(|@Y>4e>r&y9@8fvNY;vE&-8X93h0U{%`i) zY{^k%S=4@-WA>km)y6Dw35Z#Mn1ybRF$pox0_o|0Co9#RRbAby=V9+k>#od1BtwLU zhlhu|e+L;#pd3@?s&<1BH{nAGRRND4x_qrXi&8ZL9;NGr?PBBwJajHhmQXQ$cC{P+ zfXx;}vil0sR@=@~{6h)?%!lPv`r8IlwF(jwLjhp;lif{#%Gme#nyZ)CdDvP!za`&> zMmXf=)_qkH2jG?+Ti9VMRNT73LuJl0B@vx)}SJ=6u>U~{>LA#V3lNer`K7j}IKo|z2 ztD(8>uC8S43arQUYX&Gzja;nka8e-tV1zn8O}|az))35%dzRoyX-F}==c^Q>nfa`P zXxxJr^_d?vu;n}~iW(!i6Y*U%%N-!Z<`4(Gj16q4vJm1TI!onx^9U&Nn~dNlsf@>H z5>y32BqZnT*Wo0#b4S(frL$q0XzktD^xz?lR9mm@y`KtY2^*j^n5(58fzefTBXf5r ztQ^1CIg-n+St1Wrf?er5JL_bm!fscz`rsg z(0X-Uh?o5Vd~9lk$dJo(zdE`et91gFKhjU?6RLQtupaKqnzOfuja(9PRl&w$Y4<$) zV0>&76%%yht)HDIls@I5OwgL=YpJ`D;xH=^g1qSD_n0~bw{0W31TH0%IR)@o3~CsI zOV#+sJ!COX0D2S#@(EsL*9ocmkRd_8jB)_J#ro`!Tlems-fa8C&c{_~so@40D)0mg z=R4QRIfM@=$+*A)uBc7_C_-@)`W8V+Cji&wwL4^osDOeM5ETdK>^6<~w!^!zl7-s| zYSXBdQ|}&DSSb+*=K#ER56<75zJm?ChUWvet>D~pf`1ZSAsq=m;t|x&rt*Z9Ii=er zeu2qk#uD^+RJmP=mI@C}1+>M&$`>3%gUtcuk&+k&dyi^94I;@;hzkrzC(T@Nmw>9op#B@GnK>PjI=j5swl)zaK~Ef0TtERSF42G`O9vP z9*TNk4Rj<}xJDOXVF=M0*@*cUw!Di@u{T3XfoitNPDFR2!*Z81mKAYYr*Gx0%8*T{ zF4m#ez2yQo4(`MFbBz>>G@t5oNi5u{I?D^7%V(&KEN)h1VlA}{TfFT zO2Ai!voFeW&l+WKKv7u2gtA%2p@3E!QUt-CH(;|sbw@{nor7)AkULaG@A_rjm!Q#w zgpY#=^m|RgIF~(8XW~JEK}OQ{-R(nLCtP`j^h`Jv{<}aZ=&y@#-eOk&pnM+UiyQ-t z{@3L5cm~0x|B)QsC(~Kcs^VlDUyT5!-okXRkZ}x&4nM@rQ~lg-@w?jx{~6yR-$4r? z7|}tF4)nOOW0n85{`UV5}-kR`33PSiEUJb16DTN5^AH`kQd6H~7TkLYn zr_}B~_1Q9nTLXWd0r9sDH;lx-qhTuA-u6BI9wUhFHr&Q`=tBw=WEi~?>pshZqQBb; z?QAE%qV@lbB)Ydi{-NE*$B$A4UqSxRus5wu($7=z-abw8FdIUMp8xR2IQKe=Ax&6s zG5GKvT-!b1UD_-BGnVNTgHL*UxmVI&d$ZjhqaX-k-=2sH;kz5m&R@$;ddpY*;h8Wb z8(tRy`(e9g_ZIQL-Sz#zcE)?k3ml|BoduDs35@L@QrR$MYTIhcyU!2CZPR>;@c-e0 z;g%=(`Dxb{s%d__*!C1QBpkdqz=vmhIK`uh~&R|+!mE+Q}e zN#WnC)DQX-$B6H?s?1xPd{m(LyWsqp+Vj^*{8r29M}qf7S8lmS+cb8qa=@>B%Gmq# zSMO>lj9v=m{r&!ij=|9HzQ`?m9qTd&0PaDPku{*==A zeVTaFHk%#t!Fc)H{D6#dA>UHi-}1db-RFy}#a}RtKFQ5*7)Kv^txw$ae$qD^2#_r^Ao>_4V{W zWWT>(65oBdfAzxez2sK#07M+czr#O%Z56V%@%8I2|7KnOBXhOCI@6KhRQRSu{)%c1 zpW68=BjTG9c^`k4u^{p{r+{VhiAZ#I)&f{lNN ze!1fKBo`m-s8^wrk9O4S*h>Tb8Gf1{M#~p#>#z1T^y`0{js6u24RnuxhY;VF=Ar5R zrx$zI3@-!rWB-6fgX})oAMZ`^S$~oHSJkS2PM5u#djIs{3EA**K5gv}68Q^9>rad5 zZ&=g34_5qh-zvKQent907_j|!)cc73*#936|8x5PYfh;z)bn4q3jISn>_dON#_Pw< zcpdw;2>tJs`d*X3w?*jNdJ^p1xBtFUknDf@BJ`^x^cT1nzp0STfcPcT&QFTg7weyI zeyktYRQz%U_iYLK<2L%=xCH$LPO^W^67;JY_-(QAOSH>RY(Fn%v2 z2)h^CPW{^ryzPGQGi1Mhikra?P=WYwTra)$9^Y5a1oqwj-D~$jwD#y%cT!;yzI&I$ zzem8R{~Y7XVK>-0chjND#o>bd=NGJ(UgGnU6(3(G{7U%>fdTq8@!E=?{(2GmLI&SG zH(v(+eZEcjXDA;O`F^qdvqJJJhx}3h(C=F7?-7r$c{|?&0)9^Yd?@$*zK@%PJu+`pZW6 zOt}B90H$aQVEP#E`YnWMi*tPKDZDh}YZy@buf{OF2It2C`?m%#eO>V0t$@!2?+Yj8 zzhYm30{iC`m+;;*)VufOgBC?$ZU42%G7OLbuh`y~GWq-Wd^YON))zn5Ie*ydYR5Aa zzZtl{euM7^d*^rVAfRtA{PW892QSX|W&FP%oAl@V>e@L$*U=r8E{zt`#en}_L#zWCbo_&WC2*t4(c;eEE) zZP^)rdi3HKH~r6hU%wR3Z#H0GTJcZaM5hN`s#_Vl*v!V z#m}AiW{v!Y<<~DZ*nY{-{o}oBGJVDt^67C@lk)gWaWpB{-{6xqJI7}hxaeJ_D|2o@B4r6#{Mg( z_jfOi{%+s?>qY()>H4HS-u2JV%hwyX@x1L~|DbY$?bl6fy=k1cE%onz+xlVq>@Si_ zNY#G>qW;YM^vQkxON8v-rGH)%^k3rV{j(D}zG{N}W{=VPLSXl7B|83+ zzW<-F^$sz{U-bR=Hr&6j^Z&|(hW`njPki3>#JdycU)S-kaQL4a^{?E7ynd@j_UKA^102~i|c@w@fuu&jj3wr>dyvSy;$eNkNF0sOy znR;MN{lEa&o-J*FdAX9+UZoon!U>te5sl2rqr5J$dWgH@fm!u~iAWE?zBt+A#1T*Z zu+@`)ySiDz6&JWtrXQ^AAzRB$1G3id`+vLo!Y~%c95TiLJiIM4%#CM(4-bwiVQ(@L ze9V|Gye}LzdHCh^#dFMJ3h>4s4 ziEO)_;|ZScp|^`)>*dA2G@)4aF>awrlMv~-Kip(ryb-yQcdAnbvNT&ET;rav>%D0SX)FOtnysteWry($TO zuPZJhRG9`CbJ*xea=zfHR8d)L@V6k8a=BK^&TxT>ajum*Y7g~i<2ku#5}h%<0dmu+B-5y&j0Q@_f6+>G3j(IWh>B+<0U;<`EbROkzjf>q@ezPn&cTHu}4Wb)lyqjf{75a1(3KXG}i^H zwvHP*NYISRkD7(n4Rf#AYayA9a4{@&zB~xObD<4-8W3r*Z%K6rP(Ai(T?R5)hAJiu z%O-E<`*_-zVMA5%@v&^W)n5hTbjQqU$tBU`-hjD8R^|m(>!w2i)0d#*I&{ZXIa8PZ z6w-y#0~!ooPo%Az*MxY`pNQK-?H|xq3vjt^iFX3>-t&g0`GTh_9Qzfee(}evS6`< zr4J_%4Iels$+WvaO6eKbV^}L>QXuikB&&2qfuI_kZoU;uHeof$<3yo;y%$oRTw=?% z00ki67uaqBR$KtwvGB6h%BLL>*y8Y!aAw`75hqX!+F8`*&Nnk0y)~WNp0JZA7N-p@ z3WpeO;NCvE^+=#hT@;d=&@Avjd<;+r^XgrUr&u5SG$n=fF1T5%}OvAInR$qB2lPAEnLnD;i8n^Ia+U)VuMB;c9Dsp9iw&@rLY<0-=7}i_eLXw?o-XE1h&|8dLtc2v z!+|7wZ*0QWpXa`j_G(^Vat*FJQp@c+pNbQi!#lAalssn7bB)?iJKqVotyym8lbzCMGl*`}>X}&KVLNTFc}!{lLZ7S3K%MD~2*E?^hs6OraxnSg zzR}}nblIl45uKlxtn&S1r0IYvb%0`$2%b~aeHrDBa@?hh?wG~pL7u~tNdx+!eZ=lL z*rfvcSeWN70m6YH;*3nV;+WdhrQz0MSXdvhUC|qzf-EsU-d)3i-0(58K4GEeq1=sG zPU&Q5{+Iood<#kf#M($Wl zH{dD(n%b*jgcbb!SUXf6$cLA}!?A$Gv9)`M%B~(p3J+*SyX(*R_G=r4 zff;t*;bPco_%ON$7v8Ni{58Q|P&`Y-9duQp1KMj_(u0vk^}17Cbs&ohe&Y-)456q! z)_B}sHmOP1W;V7-aME%lw?p18VI~c{){mo3=DHzDwKIVD&RTjhWAW}SSYj5t>H2o{ z!}i?z^K>JZtA0BtNzojyj$hWk$^x=!E`@tUox>K?lz_IRBC)zZ@Fu2@K#<}p_uU;vr09Qe&R$VAjA;bpnjqzD|tUAUXu zn|?Bm#zA^yw0-ITt$hg^3urXv8A}iTe(}C7{Y_v z#${R;33|2aD!WVuz|7;sQx=C-$^0SgLL1od?fzQP2`=fzG+oLfxB>uI*_vX{OyW~J zjEo(VFjrIuer-A#_+nfv!&4KEHF4EIVmean$&(^jHg%JuYp}g7mEQV(H0SKoF`~~xLM_{MkM7BW(T^yXA%}rF*tqoZ&+z%Vk zIBqps!so6R!4A|0a>s+EbO~BdjI+a?LYu`d;desvYoMR5EU91eIB~3OmWPXD0Njvf z0OV4U{hnUPr;OpF2Ke(k2W)D<{;lEU-XJ?1zZe$`*uT0E+p_5e8ET&g?}20~(b%eP zKed=~JKvt4p_X0%e^%d4*E5ZkyfbC{F?|&1fE*sgF4aBBK!Qj3pv5srnP3GPj>G>ZfjVw+$@JFFh%*V@h&$ zdC;ytK(s=?yQPy`Tr7n_PhP)ifbrkGa~ zY3Uwpm3od7Y9$ksgU4Dut46op2|H`qciJsT?p|Y_u6>jC-a14pu+sUfi9V6JCE>gr zKv!vXLx-UYjOV;fnO+b7vho<-QjoE#xt>t{GT7uvFsIWTwUC*O^R&(Gyd5vkShvS> zDh@%)i_v#WRNw9PP5wn4(K+U`NfdPaV$58$~wL$um*I>)|Pj*jaZXJamBP58iT; z!y-v8K+mMBqKD5)fx!%1(C`^3$&jH)94%+GFZBm!-f`)?1Dg&=L^9ozVOcpl}YI@j10K|}75+0j*q<`@SlddEf zTSB?vdP<+8JCB}}@ED{R;baszF2slBXF0=J#H$E4*OD>#5@{{yCVYNDtzD# zO=a6)2D+%46^4rxW_<_YbV{Jzqh9FNt&N40?3M+WM58axA`C(i>42`aKAnf)UQUEJ ziQ)tt=Sb)q;aStX_N?ZvBO)n4ND!Yz6N|?j2SW=~f~$_kONU;jGIxY<_K|YG&)2p0 zgN9-Ht35GuV#kpH@GCon+aq-2=Wy^dc&f zOIv-A0doQPt#bg&qPw_&qh~#9>hvpEU}zrWNi(^$5U%@O-sC0I0l2GkPB&{&gixr` zTD$ot2|(>EEazbMD0Kgl7%gGXML(oK5d!#BdCSoB@W3|n8Yy;VbEPD7$O+|K1i?w_ z&R(dJzIPG9YTRFWqzkTiyG>!m!iUDmDfCFPLaUW%fbnVz(_E{2kXHH}#Wz^xJ*jfA z4#&&t>AO|jp~Us%HTtuG9yuO@ROjl}@^kedd(iV7IF;P7=e_#*C{F&|a}lnSA%Dlf zHN;NGO_H8F4>jNEneyf#$hnBrdqn)~6aM*`AecHmX|fM^FHGYoTO)HOMu0cWwY5 zXpsoM0T0Eu#mGV`2^;00C9R(8K`1p7*}1X23>E z&vhNl(yp)W=q2^Z2(as6d$;&KK$DTc*3A|=TrJAxG`PoYT}OMJa!v1zyhI5GAosQ$ z0QY`;hQoB+Adu}PSeXfTG3;r4ujtI| zDdS|#vcF^Z#)sV7_t1}_xN$Ob>9lK!>$V`Hu9Wu} zM|Whghx{-M(Ys|Cx(3tPp3`0FqGG2W0Sl<_7g0wweVeV*X9LXVi|~fZ^k>EO0E7t- zf;d+VzRMm=wj0N?Y$Yh!K~i7NxWwLkOEICm1O zt}g(;4c1CaY@%bG)z;2R&ba$P0x31yM_5!RIgqy2^Uh8;M>AbK0Y8YEXP9E|kD~;P zJ-=(2eVMPl;TxLgdhO*0xS81;TFgM-Y8=arbjWw}4+f3dOv`g@%7{#j48j*50rZ%dDzGW3Y=89jhAg=1_gi4WxR z0m$};(xV%%n^xrXplTdFNoO?mHwnMjgEtX%1zu9JmsY8bu?CL@pdy}F$;{`}N9OaD zcbyr;);TL&WK^qTQOYGfj=_U=PuohZF;n%ZM-LbX_uvMw(%>x&Ml#CRCP z!ZRk##uMl1Qe8g#4!D*&)I%7qV^l^hT|kNIBT0Iq&LOkrU2c=26gLxGA)ZZ@Ay$P- zyVp*Bu3el2=B+aB#0H#Y6Px+l0V>cF>bVz1>T!}+BYKpx4;6#1?v#e- zNYQOTpIZ~vAbl;Z0PIVrCV@<}zcF=k@)I9zW}4;|%b#CRNw#H+@UD}A-+h66ufD%Av<5^)^bT@5di zxB0Bno>U%UBAIUK>Fl-~l zO+pDe=S=c88t!5z0Z_W0yGkX2(~h3A5d=)gIjpQsyLnR&!DhPS4jq!oE)4Zi;9$Sw z*raHQ#~IpnDM1zIV|z7Ebo+*|$#MLHQIIZoyhLWW+pMl1p;24I=-@>y>E1D+|3*3juZ z9~E!hEql_u%H$hh-SvnkEX%<^>zxjJvbYbI@_c$c4i%_N_BK$yP3K|)Qz`|UBYoPP z{T!bEk*P~Pwm#O<;9$8aZEfvBhe&_ZE)?j_dY5^ zc^-VNI(ni|H<$7fu}|-z?)eOxV;xKhc5=_U#UoC)){`?Wv(g9YS!e58Y+!3<%qppd zhU@4vt31KXMw^RO7#1a#N(KmEFu|PejLd15YZ2hw&)c%Y8+Lis?jHT~0FF#TQUotp znLF2nr_R4cHCD(M(o>+Fo+zH|(E`zq;1d_ic2BR_L^h%aQ1C9V0bVRlFpb{4ScPX5 z>^`1*E#)Pm9%Xen?qu>ZapKPMcnoii>-p~W++H1R@N(RdVbg^OzmIH(BKwPy#Zn7$ zjoBR~Ww2RL+|UDgyFUxCQzzZ^V1dlW*Cmtc*|ekV5;SN}MtI5z=6Kpof!!4+YmmTC z+RAYh$NIL45a-*pnd;Y!zc9PY@Z?Z%Qi}3PIuxx+aR@4S+%!Lg&G-Q)y2088U3$kE z#d>WYWd<=+5rv;)D1s=4$g{yd^!mAm7Vs`pzC z%P#+%oLTS~4=1xpKuSXX49@K7fP@&D>WK6_=x4t2gF^D}r=3T7vV8C`BZK}zWu~+O z7$5>OWJ7Sp($U5c6sC)tBH8e|ti8}Zl^-p8G{L0kZDX@C!qO;ILaH9yGcF7##8f5U zFe7xe$x*ZiIz3V;0yA#8%XgY&H~Y!+kLTso(^#@>I}pm7CrJa+E`fF;1iTdbnqR!B zKeUSx;=9N#P7LD2uo+#7Lww3qeAIWMRaEMcNbIE$oRQ^$pCO9%*Eghc-o%V-+ly^F?%t&H5PZF>>XmF}h1Z zLqvihv*h$V$TS5r9CKB>!HApip@gb{M-N@TR-Q$v8Uc^e^}=>B@&X<@7bZ)g@@H4O z;SbnsK_t7c@N{WAPw@{a2rwU(Q|WITNYyGxOq@m~{K@Vnc&mJmueo}OorkT(^X>r0 zL*P^A=GJ{x5(nUx9b4F8D^%RN!9?fsr8SELP1RtNU(H=bAMDtc_y_nTx+MlS`C)Dk z-VjipJj>g&PEL=5R;owh1bp}0E#|OV=JT^fSydpF-c~NVZYlZ5SZOsw9?wn88jf7r zlQJNYJ*XSmlx1T!hRcD%J@ROac;INbrkHkeP7I2s<8XFzJte?wp$kGo(g$ zd&)X6n=n+~g^pct0Kmo0X9^JSJ-T;1EGfekcJ8QpUsqw!E+xk#hS#RgVj><0!(enZ zG}qnLm5dz#yD|Nm3H?qZ7b`oQ6o@|PefmFvwT1pAbX;3lbz$7m8% z1wkYv=j_+vBm(u3s@qFv!!*&_J7D&Jhcr@cy|(v$DwHK`fYM;DmU;w6SJ92k-JP&< z?5w+O8L&OONpuyuku6#t#nj-Q5rNjL z>q5Nj58z`{D?}C^r~B2>^;oSFu>6sJQlFX;tFRvK%bK&dhmBnF&}3j^v9x=heK0<@ ziHZrj@z&1{0PIEH8>2PP*HU*Q#bH({?Y!vZ_n0~bw{0W3EGjjYI(KnmP{SBps>X%w z8^#Ib0mnc-!K>^#AypqTB4#2lqpB-}R-rduiZJ*frxC$*b+(rfLdhq)D&b4w5 z;iDMD1rBgUZTd&CrYiat?Pkkw@w|42><|?gg+f#uoU_|B;@b|dy$SX0M0$-{IrZ*g zg_RP4a1Q&i@!hI~F?+X~JtC-^7P71G(9ay)|C*;JmeGN*LA#4j+J%vgdR zk1Dq-(Nf{TseravSowluXs|h;JW>+FVDC}Q=V8P}_POS>f~>PAB$vzMDl;b%CD2f> z-6G&`E85{TXvAacJbZG8>}?q1xv@siGKsG@JGI44%^CW%GVJEx62he8#6a~)g;Lef zthEtBq8<$K_F-}k{9fjK)5A*OQ2HLTtZ9Xeg=$T^jbJ8=Focfs+-m189ME?xL1 z!>RA-JtqmvCZ^LanRuphIGmC8=1zyJ*>~<(YrRu?oqe@R*qy)Z_UNIg2i8DGf`w~z z0TzZ3t&xqGe__kJ=oEW1q!h;$lAVa|M2F=rXDln?v`*;|+M^8Fgz92l5?jkXJ!iNN z1QR!_GO?CghOJ>N^}#(47ctyK4l7Uu&b}ziJ!_P` zah!F+yyA^lle-)joSx$J>D z6Auy$GLp9MZXen@;mT`5Lr-m1=LJGRW2k-&rEE8%bl!~V^#1`VV-F$=lMaG!u%%Bw zV!Q9C{ih!RPYyr)-!1>|L*?vGdA?gjYWga_^63!lB|(1h&*bI)-t11OMhy=Oqx^fo z9Zr7tl}O^R)hVXX@5?^^b5aPl=HFiD=Yom1hW<-Sr6X^}tUq-456L4?K;^g8x&Bxx zJq5MY-=fvOMXSSl7>sT3o_vc|{}!$OEn59swEDMb^>5MY z-=fvOMXP^{R{s{Q{w-SlTeSMOXm!|6_!h1HEn59swEDMb^>5MY-=fvOMXP^{R{s{Q z{w-SlTeSMOX!UQ=>ffT(zeTHmi&p;@t^O@q{adv9w`lcm(dyr#)qh*G`g>sX-w$Jp9KMbdccnnz%NYb-xB`&Rmjf%JzV}D6<>}~9})_-fb#!6@#R22Kp@fR zLk0sJ`3j_Jk9Lv`$J_6NUB>J%|A4W0Y#|q z=y#3v;-J0jfe#7{r@lkqd+wtS+hcDzL4R8o!&lW928mSvO~leB0j7PsK>AT!h2xmV z&_TDUuHe8X0H1??KV9!w-Pq5%qn?pj^+Xq`7+3&mS?p@qoGWq2#KVcnz!!7@AaOW!<=;vnH4eZdp zs|@K~BtJE5+%^sH8p5bbKc&vS_r!lmu=_*B^F5D`pg*6S$RDc>bd#!0P(F$;`jX#& z&a6bhGyXS~5zoUtP2M}>gZ^mR2D;*>l?R+EFW-^S#!KpB+MM)<3r+Jx9Poc>)Bt@KcBAh zx>miut1#q!RAJP2)W@$EH|i@*_Un}FkMH%qyf!*;+wUFwv9teq$av!Y!GEr^|K+6~c*bT} z;P`E|&EUTbie00MV0zZcE!xZ%g3PPEM~qeXC}*C>f?s5gwJ+y@k`Mgj_~fzUvqa#S zmH5IKo-I;7Aej!R5?Rov2%{0e>;XiKp#ka#rTqkYWXy$;*({LQ%#{ItC~;;fbJ^k^ z09nW_%Z0ybCxHd<;YuXpZP}QCv+`i_Bd5X18LryPqB$m_v-`$y5JD@KS$mpbG<~`{ z5AefMgDV9U%)$nHl@clpgSX`5T86uAdK)ZnFeIV4aB%T z^2#s>Z!+B5{WC^jG8alVY|uUUG6B5z;B5^!j4VI}U7vpd)(oCST+Iodr#iBG%Pk0J zdnGJmG-#0x2YG-i!V%`g`Y3`k>(jkHNFw%QoBEa#Cj?(gfsfE4F>*@B@oqcERp!?*zwPFGpSu zyks9k@lmW@2Cl%Adw)7eLT@BGn+uq{;GGB9I3ty`N5>p&g}NnM)zAwKl+yWlzC3c_5|rkEBf%XS z=^wdJyy4xA9kml`CG?$9@OODZ}5Xl!XYgK_yEVzLhowds{xAiR|9hCx})dI zxpzL)P}gK)YmBj;P-5*fafkx@83NJS*M8}B*Y&K)c3 zc!|U;tWF%Q`2nv67El8{)=lN}rf#&i8e9obx-bsANBqO1i*l7!S))2&(V*4oQuYEz zFDlld?b(2uZcrC<)kYQ(Rh}6h_X;Wp!iApFlREIKyyyIqK(xi4Oc`o|jm{Gdv{1MY zw;;Ywm)OqbAW@lCak%SUoQ4vr*AM*@1s1f%%1P-0EQ19J$3yk01B}W-s*F@ z=JVh~Rq=oi6E@bvZieo2v7irP!_Df%@RfsMD!1){+2K5A;SEPts09f1r8V`uvlS*6 zIcSvFveZTmwIeJ9pv1MwVJJ1`aGosGfqAnvL#@OMi2g%0q9PF6=-Qv#w`XwHO66(* zF=WxL2z+47pyCd50g!N81E|n@(V<9bk51hVY9FsHldbj@gOZ1)4^$&4Jq{|2oh8M9 zS2|5lksCh{HnPl5t2ysiz!mst8sSjCZqaUL0YDxrtHMg00pX&bP6I_~N(}`DgklmHiok z@^3=D4!6<vunhq_Js1Hw- z^IjcSuZq&9zNY0V6t~N*F!Aj+4L6`mqo}#FZ?}qt_#NnvC#Z=xs`c=at8ZEl@c9_wdjoKI!cFqF=7faMV@ z|55^6mfaeJrYYn0 znxhu~aNQ1T*UhjKhG!(Wu<;=fd-5o0Wp_a72GkHc0QypKOLYfDnW^mV;1BT`G<;`T zvBy6K*{#-nD@!s$xcsM;d<&p^( zR_$H5X)$zN^tgk%N*@oqoT_l*Y&E&c4Hg2vObp6-cP3E--VLfpN}bFq*E{dOvAa&& z3l%O@_;qu+>=ofqU*L!}Kvi*7@30$R%-IEK&r9q{XZ*aZ%CU8LKXV_@$G|=WQ4mNf zC^TJp0{X-URs;YEcJWHsOBRIjDwraZ5~SoTWP}!6Neivx8iHWBhY=N^>^Y85=l!$w z)h;;kjOcp$SxzX!7(h0YijA|9-FFNl-Z#XlAZX`+LvjyFv723Bt85cc&`?;5Y8=9< zoQ*gN;2dB+xmd1dMmzr=J@!Z5fLhVH>~GM|5PlNPG>4~Fb@5|+*7MDN` z#HYf7bsR5rJ$^H2yTR}^)~lZ<$gc`tgt&;Y?A{z?||7QSxF>l-AJs_Nk`^` zR_=r7W@XuhC|IbE%_Sz3;`Y&fgB)0SZ!VYlbOg0;3;f0A4X9RHH`zIz&TUk3D*wo0 zO^(*McgJ`TdVd?LO9^yh}e~J7OQ3sEKCE z`QYB{94fpvfVU?1_$ukL59mk9&CWuwb(P%2S9XCz3eS!tH_3Ds-}+~_8EE|i519>`^Y)kc3`DknK83A@ZvC$ z`}KhzY}Wz>F0f__^|!OLNpp_%L`ZZoa*Le3oHV;gVFGOtbZmcRkX1I0q9kxc43&hd zimah{Ae8wRa2)NjMz~#%*WHcfA2)zHGybzPRkC=TW@7N$C@)vVDoH|S^AcqhSvIEo zb;$3l5#DOmGpoZaLvGJI7VuY-&Q2IBC2V3lxeUE*G~8 zv&dkiZQ?x<_X; zpgF2zQC_Y4_6eKTtd6QXvG$!KaHa6X7MPiF<}?5nAXC9boleyrZl{*h*@IvE>w$kJ zJJn70P>-3$Km))`80X?$JGhlvbN4eAGuV14G4BD2k{_6$YIJNcOK2FO>2AGjJE98& zIK&p?cuuZS!Nc5tvtbD;c4io7MzD!@5^i7=4J%P6;ZpWsy!Uy# z$rgr*XlSW4zoU+v(u*uWJh_K`q+q&@cEn&Aa<+QS@h~UM^`6(Zje3$6vdy)8Zqjw- zEuYHc^ByX<=F3VqV%0F=ZJWdmom+veiwyyBFnPYpo&gEu5gxl94sH?vFxR?M3-zGx zHI!neW=-!&)Kw#7eIhd3Jy(tc@SAtvI5w@wCrbkCm1eIcCKgT%Io*;o02Tr1w>Ptj zJ0cQF1!2v{8&b(P_mjOlkl_ySRofbp5u_A~e#1|W z#53M}JhiY`OpwfA7+NKEF_|U@6Heai!Hwtd&=uyPy$0*Wl0agDSiYDmtD2wRh#O(#$W)T@mEobT_}%rck2{JkCU3c{YRdy$cFMwCVTRY|mBXq_ zUAOIQ66)EAGM_jKc;`2mQrRa*+*~lX>!aSkm2Z|y`p^&!f8&m3 z-ldPWuAIvxVfP_dHX~d+x+>{0IS^p2-8S_MghWQc>Cv46%k3pjeU^o4+ss%gGOWJO zESX$6hxXJbw_iOd<_s6{tXPG}l>zM!cBN}8e;a3wNRC=o9rLH&p4gkyDuSb_N5R6% zlW}f)S#4le>8V*<#myp+%k|i* za)pIxffJz>h+AEb>2TZ#68`489KWhUMeqLSS z=!VD<5sVI?PbZ{XOS&;gL1pYgLroHC(QQRVVqr9u3%PQD_zL=%ndaS2s-QIaljZHC z&c|kKV+FI9qZ@c#A6+wIN~*!Z5)hd<2OfX|E5@2RE5MwJU7E5j`KD>gZ6p8enX(qo z@?iw*;t{=H(@`1aY@`7J_Zx~YwsqxPDEDr~ zcqASf)^(-(n+pqpg`oiMiglT6&`|8e4X0r+i|2mt-^A4CL%wT!%M<;(df#3%i`_OS z+yvFjT)zXyp4Rw$AKsm_zjgA3>$a6v<*(k{8o0iP$8#Aqk#Xh@8jPj)x+1%_$%URV z3w^`_H?|TtwLAJ@Zwm$>A()+ZOzG?WX13j^RV-tgVthsyg|)!)zSHtSsvZVxzw*||1nSXaq*YcOtkR#vV7 z3XNAOtScBBjDfx0cjr-1O|MIk(pLszRv8!sYy$s#9@vcd1G$q|)Fr0H_iqZjqeNY| zc30dHH0xxKz3O+GqXFsq^q^4^Us@@Nv%D_!GfpP4mUo<2--3LFKYDv}!}NC@u=ENG zCc0gE@FyR18fotsfW%SLZr`k;J00W;cFK)AtgUN%IS44v)+kU`QF=dnc-i}S5m`e5vj*#FPcbuBrn#lT-OCO?vM&VIv~WU|S8{S~vfYHD`J0b@8vq17$9fFP{G z>^aZA$FM$miEQP}#{!ddE)sT1+l>A`RUfVhwPhf6q)00ZD zUn0beTs^uB=V**11x=QG?sLx&GYX0`OXTq{HZ@QiLz>O-dV2B zX+KMgCfi-i3p-qI0X;e1<&n$^3&ZWnHNBoRS$GCa5bNiLA4S*7_MqeeW8`+M_J|^r)Kj4*C`ZD?H0mQ^KQ57P4q1eS2%+Po&JG&pLh5*j6#1iyzu*}sHncOP)p#i7Yqr$R4LZ>(S1H~ab^z~220aX zm=F4$6izUB2qQ$1126>E<$4U(2KQ;JEv^{sYO1}Yn%Z$QJfmdcWUm-(7xCO1r5H!t z3nrC~BqAJi1cvBhZSPE)^|TUqfE6J!YHWJ9W4EV$yMEqTCj+NVEU&0ASG+^VcZ)S! zw707g8nMX*VcRt39c98);~vFxc2~YW;N7a4$l>&Tp)^8Rv-=5|ZQXz8`}vWiPVvr0 z4VImY*-Xd@1VxlXXll1?%vcP*c!QU20Oks97c$@cd5_qH%zN0KlA63-p%mI(XVCHX zC-Bjl2gpm})uR@+X8C5?r){H408duuq?~+b=vr$fZGGUvlP*4_u?D`b20Iu4MnZ6V zD(=usa)JuRk~E`t@qvWMlfK#j}PO{g4#{=FzG^$XG+ft<{x#mW@mjV1Z(>aiuROJA?kkYYc)0k>8cA*0O4F8jXwi z*)L#rE4(ll#urQac*%?^T-z^^m&ZpXxzP(Nn0xA&0}KW&p7d;cP^NZ7MYmrYL|S4m zq~c(_G;YtB(z;-d?7Qd-9+`rwRU%H99pt#AwJ+hd*1$!b8MW$-Z)ub#K_MuS1+uRSJw3g9^aE&ln3qdyLl|%BHX`{sx0uToY>DcgF6u5J-pe zT8qY{KN{;WfK8;g4ZLlMq|YzuJ^UbW4^;Cs6!~l?wu4nf!z4($ib|LudR!H%3*9gM z7*bk6<|9(W%-`Al3IGb@$nyci026!WQY%g}Dr2J~=}JVwlO9Sh0YY8~%zPRcMdZ8# zMhIaJR}R=3J&y~LAOMmCMUygsY_^j&?md{wMBRy>>G4#4vCOsd%b&@|TTnjJU|vbM z4*?6(^*<3}0bkrU6W>gAhD{JikON#fe|fAG|8+q{pku=}ZeP{8PV?t$LfE zJ`UnJ_h=CoqAKy=&KCb(CIe$$6qJWV zh2VW|jgm6HWM5zd7>y}2y^;N3G;B3x7T1YP+fDqBqzEiIMl>G1e(NjO`7pe5;{gXV zxmIlo*UL%od?VbC5s%*I$xsV^MoRat!4#|n@3KW9@C4K zvS@)T&hquyX_fZilo7#yO`+wmD2;!-I@>2YI+`%D9xP-BweNjkEwlYSr*K!aKz+F{eS-O{KuZBv zaKXg!^!$Z`qzB%nHJXIx3p^n|nWOO4{lSRuikZfVY2MU!AeN+Kg>hyN;dqSRrgwaM z4(Pq;?i5yb8|{KzO--17`$*}P6W9&p&oGMeG0YGlm*ROm7&C?k`{$DkJUkdmTqsqp zf9>}5Bj22G2SUo?l1Ryqmr0~~B=6MW4Xk~IVS)Zda_27#N0>!+HlkmZsn8UD>~mu< zHZ*FmVEc*%u!Jy0w3qhb%yh$Y8e+(xEGsgSdY6K73qr%7ObFey$&g7XJf>G`PRSk*d@~;I(wWs~#hghm(bQrn09;YKo z43p@JA6Pb#Jyd{rzAy!~$!g4AdzhCzxRjv?Dqab>#J?bLa6&jT{2=a9ehwadhR1%D zptku0F6bWHcD{}ATU-pEBJ{`ZjLn#|4m;62^85X06AE6@$ki3XNZu+0z*w+io+Q(NT;zSyWoF>AN!NRj>AseIns{ch6IjR&0-pg zE%_4%*Wvf)+}~Kq9f3CSr^uIOC$&MBRP_Mp@@aTi?-YR2i+_<;4y1`( zWF2NES-KG!6WAN3ZDpt1@f#NY)k8V4rwXHi`xuRSo5x!RMpmhSA38~gJ zFnREr%})WfdxATXg!n=nAgG7?DSavu28JMDXxgf0(%YM>Rt?Wp0R5sFU5dyb^|jyr-zc)9QvLNk$Ya){pXs?frR!;SQ@;unwO5pi*l}jNr1E>#A1ys@E8^0e(PKDG0Gy zIf#z*l{Wp;_vYBjrkbfj=`q7o?HWJy*&Vqd%f2`SH??&y+O1D}TdXoo5jBN>kFWK2 z-#>_!qV(;A`vYckx>3*+2Ckw7xkQ~7b*_2O;aY<;=8$G&trmtRdv~}8NizeXXpTuN z#PqP{EsZuHvh<9)Zk==9#c1>!51reKXtat3&Xc=`(j zg|#r&EM(o8rKh>b<%^!&YZG$)fMLh5f8OB{S`l@x@ZA)OyAQkky7q&SKVHNWD@ok~ zODNX9goyLzmS`QGyKI8V`!+gPMQ~b|V_SJc$$aEl5`0N$#X5V$_iY5w{|%*O9&We5 zL^)YF)#V>M&%m2rZ56y8zt7{x3vc*{dq^P^YG1W9sm}%myUNKp+*&BsfoC_Fi))b^ zw8%Ser~i@Qv>?NTu;h-ay--h9q%3 zDoHU;4-14VSm-qf>Ad&7*rfS5WW@(a3f(8+jpDW^r++Zt?SgiI{ZS zvQE%x-wkSr!JWc7MBm|f!vP5XoNkX&=US4&SW8K@GFfwn@h6`OlH|VnOvMl8{E!i* z1Mh?_lb&BGwK43-)+yVsvw9q&Io3uK$`J-w=WzJX_BbN=cUJn3_BEQ#X^FtH#E1dq zpY6-xdu^UGNoUsC`iu?ShLgO@gXoSCutb}Onl&9={HhRc1v6x@IIf(>QK0c zqHS|Or^g#35WwoNFy-+)+>n$6wr7BtgH#URe zJq`1bL}F$Y;m&>B@!Z=d^8*}+lcjIJvuU-jmAISVjDW)8@9hXFh7)m0OCv<(ciIHI zZ$fa2$}s6FACNfUU7#5Z)Y6HiTEK5ENmG|_yh3+KU@09}47?Qp_c@p|Pgt@fFggyD zrO_Gz=P$no{wm*JDrryU_VXXtY6~4DS#WT?*Ht3rD;nh8j`ar8UzUuezkIk+6B(Rg z;+c;Jd%URG+vojaZ>9rG_Y;3MZTkbkwP=teUZuRQpacMHEm#FQ0n1t)3`@Y|WNx?h zkLEi1;j?Ws8m^L(#`Wf#LI;Ah!At#^i!95*1av{#GbrVh#CNPYp_7H62HP{e&i1@e zxwh}!!3kjsxEa)tb**Gd)_HELJ1$tlKN-}2xDLW)rOzc)=!RTvbCHLe`zPH(OF5F@<&C4oFf_o16`Ra&4FP)vthR&!F3>;`7-Uz&KeGxqXg!jcM3? zA&?cHg2x3wAP~F!nA4QHIVxgJ;Cm+mg2RQO9uySwMOS9zRiEelR*YV^VU5o@mXx}g z$U#+ShsBGc_kkDaza+oQ#MANc8?}=k$BQ(zmv4R&onfFDaozAhlt{`8VpK-xcGy0U z3Y>R1p>sH3??J!L{B1k$44XwQ-FZLF3LwUMsK$AoyJ|@D^$(?B_6Ljy6)FIM8=Cea zDB#05%zelsW7)O96(VJUjf_A1yE6HO3E;!^)woz82L|A~`&ux{9=q-3}kbjC+0Z}MsVQIhz4dB#sw!B2je zG}>_RBqE|CM*_8Q?eYQRX`a(<*|sYDsTrUj>PP~6?B&|G#nCC1oF0lN`aE#_cvsqJB}OFN zQ(5m@)~XOXi4148l;PvrF4H=8r$7*DK>c73I$(`-l2G|3M!sqA&*CQ~D|a&4*971- zoZ%?C8A1rf3itJqFZMfhy`ZT&YXQ*X@qY3(uT1@1KC*I7n+VxxU_Fw{_38@rS4T?Q z+SIWYFsYr++^TjsV94-UGUr_f&Vg3WD6fjXnutkNKINg9@&V|857v_`A+e$|Bc{$7 zA<#Q4TqtKOAiD$~Lzu}<%9tkX= zfqEW#Ul_;&D(k6uWJ`eme~*;dO8S|>&fZda$TG%&IRX zpvf%+Rk76>O!@T%DD}(?f2LN8J|b6kN_7dwN!ZkXu`oN?*?vAnF#{O7-I4?NLrmacXKKI%1--mUmS@e`% z8hC^KG^MXlU=J|nBwRICrsEulCnf}w8s7sy+@|mE@}Xk2Hk-EDf&EKIzg$Ki+x!GX z=tfgM9t&4_s4oR3cv8 zLl15Ip>L%CW6OiZ4f{)IfYSEW$4lgPK@E6n`gA`Uc$WP)9bkq$g-rzO2c}w}?3;;M zW=V12J#=6bukMeCk8E&O(l#e9cT5c7hxPuRtm3)VA1j_wPbb-)y2Ey{>S4a`**x5( z^Mi>kXlyPZy|cEl zIUz;7V7%!}4}ZZ6ia~~k%wW*J@e3-8^A0xtAPF1`+!DQKLN7qB*FunsldQ1f`=GoUl3>HgH+*w^9H!b(44GG z*ufpQ;R`k_$+c}x+dc^q8iHRXpK3{&d_b1mI zcJd#D+yY@!YiH$Sw!Uo$y`^%@g4=w^E6_1j+r8rhsm%aSPnhWEuK{FW*1Vdr8?POwU?8!}ab3Qq?#dHZ&mZeIXj1ARy7fGpx21Gfb1OW;q-c{H66? zzqZPB&|m){cEedhdCPBC0>&Eyt|0a3@o-&7G)gnRVfLjc=!!B)2WZ^OvcDU41Is5? zvtW1+6D)%BX~V(E$qZ}S;Z+ZYy?ro97va6MtqXaq@d3QBL;D(6D-2`3N;PFI;_1(J z$AP{$tie>LS4qu}xf~!kSjk}P9t^B`QQ`wlVnjThl$C$~5E=V`#q$hCbzy&(pxE#o zzI)$TxcB5!AQmnSFRB^aoyf$_;0N`@ib*KdJna$>9ryrHQYrj?JyHbXDyM5Zk%wAV zMlhl>``bB4q6Ec_&m>4vr8QhPF`S1zq?INRo%;1Jx&$J&<%~d`a3D)V?Z*ZXY<~H? z3qS|Vljm!}2o`D_SFnZBH@OPQVn4kQ!X)_6QOaC0nGT2q586oKo;qQ7G1XLB5u(!Q zQ-R9VsL9qjLZNl#sk-}It4zQEbiwGw@4SjHVGEzP*mNvIT%CuTu;T=E(!eAF2ZR4#>_WJybMPK5CnQ+PO*jEpwVf2PWa~fx#=#N{xfgCU(KEs=jU@PYxqUk?pyh6C_Ogk?Oj^jnseKpw3*k;dtyk))t@*Vp2Wd`VS zH^`|Lf*v&)2DVp(yy-w{85k`Xj>|mVFlasssu!g;*#pLcGE;c%;1Oz4(0>k4Mwly)PsD=#1YedYh(D zKqiVErEI9keFM-1s3$V!+p_V;PT}no1k?EwIj#WK-EBsyD%cm-L( zmFDvfyjPY0m;zWGD&=|H?GRf#@3?}$_*mdPw*{tsz&t@AcBVfdF$V8cJZo%g>N-qU zOd3q~=N~iywwCUV#=#)oV1coIhJE(T@&bTiRvEiE)Q-C0*K+9w+ghxiGGS!|*b5MN zI1TXzYu2u!e{MlE5nlePpJyDT+u32KQec`$4Eu4R)r`~BN*M}tm)!nV;x&Y<^ii(O zgCLeAs4ar8dd^jH1vzd4+u>b^s_+%(HWpqlCSX91GMVq^9kH}o^vYibV51+VeV4Z< zYV`*tzrpQIO}RheH3?ZcSr7jcV?SuOOH~HMMQ_A_w%^IxoMtq=Ov4s5NHc}F>Vmjn^Bzs{_5APc!55a zy*!=e>dyH_@V?Q8g*Y}6g1>5OT?hQaB@u}`R}%euon{Dol}4dK{5S2we!IaAZ?W1`*A%3wI zY^_g@Z!;pYRtZsGeeiMO4;7`tK`E8CJ2A98fGifQ>IAUfBmw28+Hb>vW_dhgx4}1D zr)fw8&=$bS5Y?0KZ0iI-9I}_}dmhgE`Mm^)T-4y40R0ecxDMjC5d2UMOWr-Moz~VX zFWCpnlMj0%lgw{b0HR%>ef;1?vUDCPQHKmSUh%gU=e82|QOH)8w^#ppF zloxA_wO*x++@bw8&Y9t|U!9|d7c4{Rir_OlU8^T$G2Da${t&QzisvB<`fDFw9w#964WdU{ySCb{vZXISrdZMhRF|B9*Er(bQ~C zP&;Rq2(>xmKs@?(M34)dZwdR2wWtt;8&(kxBH?t&5IwqAq_qqZfckKz6cbfXsQAoH z#c!(SdP_vv7SuDcN#T+_X^+?Sq0L1r9w|>hW*&;301hbSgaOK1*NX(8r0`A$*OVg2It=_J%R2u--Vk_yOjvbAis1 zAjB=9aOFQ9Bi&Fcn~0$h|$G^(+b)`bBXL*2=naJ_T#xt=z_X z%~b~tNyxR-(YDlc(PZ1anASm9}zr$Svl&1hYz5WmkobhQpDZH z^veSk#(OwwL7$yENQv#=H}d{An0lU2)YML9>7m;sY`wlCw@hZYSN#RWy93ZWuZdz8 z5MB)KLnMfRNEUB5Ot;ioYYI*DcEux3q@>()Oacz4c{SgS&(ig_bnS#~d&PGH_=6!D zgHRoptO-yGB6d*sJ<~avEoNfo@d`vH|2`=Dx>Gz(047%JdWZ%9s8redh=c5fRlobr zyI#@pRpnDWa+fo?c33*^ev^~-MEAt<&bW^b**rS@G7o#fA4Aot5ED1>9SmbF9}00x zFU4L4@zF@x6Q&<#ilvU4f<-G;&k0&o^B>?(MCSmZT|s3p;2C`Z=y=@8MhtF{DZV}| zTq=O$OZXXA`|+Fv$ny2qJWA!WK+jKk=^z(}7wEU`%zPYy;7xp$LTIKk`G&Vb6>GnM zGg+0iZNGWsySBQx1Lkk>Kid8nZi_H&ndbBjtB2>Du^3w|Q0XHOF~QA)qpbF4 zjjL)lTJHfBebn0FL76tIP^k7wV7nP-ue^4ey>Imj(4vrWMG_goA0oAV>is}88~x6O9R+f`9Z}vDpf~VW<+9Zj~@h%d>;t+ zNw%vG%>lTR7f8+kU_k$1O9+8?czillU(F@}nex2jX@P*Y67&WTh|~kbZsCWdZwGkQ zh|krC?(xR6xs16>%Mxl!>%eT?y?Ypb8>5aDT);WVkZ00u#9(cRhtFExsX@t8$nYvw z1`$Mp7XzmR$wL^PfGvw<$@+HtBPV!$zd23lJ#;*a*b-P{`&E(YeJOG<Qi-OeM4CfGqdrWGm1&;8hEe+ugzOz-EV1!-VQ^H`pHbBFjHgQGLHOV5mT)0ASY` zhet^?+1%Ql{GNp_6ZeAu+}B(pi{JY?-@f4w3Pn6#BFFD;z9@dj=!cc} z%9~9{Qt@cue;yhSXoc(}S9MHM_9tb-)nngs6-Eg-7^~}n(K^{NIwY!x7xyAZ&8IaI zZ>-XvA%K7g$1xF$qYoN=a|6GQr|_8QPTq(B$iF(!y%FHm#5S@K9UZ5wf4Bj-e$Zyx zF$h&@s5tV=_BuOGGc@C(mXCezNWtxKHC=Y1Y8fx^)w=aaP>yvVHMtYovP}YthRA4exFB<07s0L;k!O9jxzI*jve2y-#RUO z->cLWTNJ!Ygr@dALR><1W>whKKAle1Shod{BrI7#bB4lmP$4=JrwuegC_(L! zl@Xwz@W~%(IE{`8q1Xyp3-t^jv4MwW2!KOaNkMCP-*QlER)fVPQoZ;Dade#~iyT8o&A6u9>Ug6W&>_#Mvzt=@rJ6ELFXY3Ve6 z`q2r@NfuL8_a2%N&=YX&;AjSXtjGWGH=nlC@e&_lUV^f$;`&S^c!PGg$Gr=5yJqYC zvfJ%ZywR})Vu9~7QR8?`OANfmJRkJm$%)2jO4-}Qv7}F-=yBpMDGDITxtNk0fcEg* z!RB}w<{f_6*?r|y{^y}(Qkfqhxj8B&b5pS#0}yuH4xRut8u=!4A^WVXTs;C%;(^97 z+MiJZvCqLk--hH1UnJ;`0(dM?RUE(2;-QCDG#{THfG5B$7YXHsI%W;dQi6Oo!!}IT zcS%t20X?V*2vG-G`LEI(K;*~JBKGlJJgD7_#7<)QR&c1v89$*m14mqvo-1x0Z(z0t z-m78}$fxk3;+d}x`T8fx@!XwI(SeM-()VWVL2>p5=LU62K`3ynjy^iC6#S$^+m}A5 zFK(FXC9RnW&|o;`P(X>SU7l)tg9f#5o{!&k)IcvsB1pgh*?mCD;LK23X1>hL?6H6M z?wWa72kH_(WvJr0x0k){z*JExzMKMtFhK?cC3?2nzvlK*R$!2DxbG0_ojb=Da0bd6 zlw%=pab{j@#8v?lL}w^SfqQ!EELqhXYmP;T?*ucOD#W#*Jq0%2))djp&Po@KL$Ux?@OxQf{eH8Z7IS}F`g8xng;_l z|L?meaUK-dG7iw++pUX2@2|sZzntbhlit9BqNpw4#+Sxo{H*88glGW{vu#9Dh;ave zhkQlczdROmC?G$%lylBAE7CjI|*s^Z=Rvg2nK*u!Q z^FTiOGA?)Z1ez<2bU?^r=S50}^7_OXJF5&Xc3%$-vHVc{WS&zC$M6cY@(X;!LL8>( zaXoUc^96`nq16=-KSuSn>*eNcF(s>?q;{+$3d~%Jk~=W5?(5OM4dcay0l|R$8k%Ei zJuVHvK^hR|YSDsyM*YwTixmB|8|8SZFLb=kBx(REC5sDTXweRTYZtUxc+agkK&u{c zB2dcF)AI3|#mC-9%ELrLEo7$!r{8oWv6&`wSg!*rTfUNR2$)nn#u&IZoicAC$w8A2 zQrH6Ck5l}KOvLMtBLg`dRB>JAGB^)mT~z#_Ixmi#vE@}=ii{WI0rdtlq3b+&jW7;p zXs>PkpmaySnkSZ0nnarg^ygE1F96)@jTluMm# zdg0R%FfQKl0saq_5pYeS7=W!+N+smJ{qAOOx*tFdhM`CC}<NGhT~fd55F6^4(hNJ+}b#ci+PXIqP#=)g*rxGO0%Q|rem zt(y``9Gvuh)gJzu;N-t=@3uyaU3IQkOTP*@Yo4S5Vx_ci1^}T>2ioTK_bC}fv!c^I zwgQ$Wd~f(w!4oX%QVT0X`!SHy!k?`G>pX3%&~( z_aW^sUhwLoF;Mv#JQp6vO^|4U>lDu?2(n@T7zQojuJrVuy_cNWc z%$WzD>^%?<#GS9CcMFo@{=TQGga5SlU~Z{|Jp3`-l_I1wFqz^5+7`hZR=urT0?vs& zaPe8KZADkt9Yjd7WO+l?n$YS=mj=2PVT_Ea!{teICXy&zu@g8fuDukr19T;tL{>y6wJ>EvP?FIGRf!nbq{D1 zGU$~|idBFMwa~g&N;atZ7k3no<9w~wog{rw1>{S+X@sa6_}vojAEK07?l6M;}%Oe zNlS3gg*4$>SqlA*;MCU~>XZavzVG;wHX6F{U{BabYrG{61{{!J404G%p)XWf9JetC z1`XGSPO!y6zWp^Q()=9?AHg81&@Jf-=;vJUpcI0-nhGUYL=OjayKkbTM~mpb?>tG@ z4&}P>@lIK#!T$VmN%Y$gO<%ARD)^KGS}7Eq{$AwltDir>E}n3`erv{Ifu-pS1E!73)y#3*;LAUuX`x!*~hkl zC_&nG%>lI}bfW0#X+f4EZlE2`WfU^7Ih2L{jyq5F#I z{QEXMVwaG(81M|=(6_-#D#tbUKuUbq3$0s4$nFomsW+>QRLH7D?tXF0K+mG;;VM=- zW;1R+D9D{7uMy19%2OzORw;eZ(z3BZ_;E_i->Uk4|dWFthP-j z=`uE%Mdteu3;l_+-e92M@7Tbjpq>!~Za77tNdELI(C4a3enHd>cHMzo8fenyTott`iDnv>Z8IAipU-^tc=fBkS=XnUsWrwFs>? zH?*xvj@DAp&+)=qN4LW?d@lhC(R7FWK_G1&hQ8mTbSEG(1jd!IN!*@7$h-9+DWA4_ z1{k`zP)C4sNPcZbwT{zf<}H8{Pt7}6oyei9Kx{lf?dpSw57fXpsL7#=R8na1blm;g z*)|n{fzCeX7xcIMJfL!A2RCe1JSI5{um>)F4{lOB>rLO+|NDko3tERzh$0rkUqhej7{V*D2BNeZ+$+B{g1vLK=Fd( zzFHS#4$cVK&fUfi9+6jM_6(sf=q9T{1tbt2sYsrU7JbHK$i&J)ffR&rGEJx-9NJ<3OT0|r0Pa41tx94tpmr4G>sIgF(KW!+hM)KBH4&Sou z1fYXwPXIEu>RGAVia%!K6VSNJaud|~#<#$We)oCUjzKdhBpI0kiU?p#vy*7QDBM{0iVQfo5bBoN7)51n9SOK z(W{{-(DwZB{k7hD9HiuD6@gl$fo7Ky<=){{t^_DRM-UeM(Ej1wMNAR-d}~kGG*MSx z&>f?>^H|h(9(0;~KZeR)sy09G5Ny2>IPn|a!OG4n4?{#v&_xWg3WarKcE-`Bt38HS zj$bzLt$-ut2P{xH8s~H6s*#}FYk#v9=KZ=tLrj@3ZxqQ{Aw)*+_d)XE^hP=ZTy59j zanSV94sYni1a%Y{I%q-X<5WJO&I?*O=ym9_OaQcHV29d1H=Q+y)uWuhH67wVXaFOj zkU6eqZ<_M$A3EW;;%^`ZW()yJ%JD68JIkgt>`Mh8jK1`rVZJD6BP3ySRpoOM`a*3iw|_y9I{{F zP`oETS`Q6ivW0ggSlSx@`GpUd-~*;#Ad;itTA*OCAb030Ui=Yc-(UDw8X9F99OzdG zG-=#Pf!MUcTiB;Q?@Oqai#~LtSiBuCbjqmawdgid6r;jDV#Dv(KeJ1t{1T5rH^-1- z81y@&w6oOQlnQ9m>B{!&#{uA|}f`;01 zrO<T zV8G^0faSYrd<%Z14R^^ta%KrMrFj)C1aB&^Nbk50gB;uvDGA6eT;L3|Wb!rh0c!1_ zF_LI9ve|urv~Ss#^tqOsYfutRjG^4PC3k>~yIAzCxBV>#3@^f90iy(8%d!+F)6(`m z3k-Za%cvyw{SKXShF;G*PIz%E+2M*b{_|`wR`{+REnSB zId@9fU&vt}zC4A6hqh0_r|YDKdBf4yW&62(1B}4FhHDA!DPriOL`t#@U3{V1(5LsA zLby-v7xr#l9t(8`OTFt5=kV}MB)7il$Nh0D%R-+l8JeXguuL3zdZKe~N!@DdzUWj| zf-tNfR{4s~De{&v=(RQ|(13;j>Z1VMu2?rQiIhlLmcbmbAui!wtZp}52%-zZY&N#l zrpGrvQQ`6Hq0Bycl@(L~h-_FpeON$C>nh7_%N18aBL|a0TapNNp;b<`H&*RuLf`UN z0$PFFc+hq%+)EC^-$iOME%q3$bap8%wFq5^6a&7p>gDbedB@}i9GWu52tNfQ32rLs zF;qZ+{XIb{$^lYL-TucApwBQXsA{KRS}8LZ1Fc$vJKfFhAOT|HQ|6G1YoW5wn$MvC zYwMtvHuuJnE7$08pM3+kd8T5kJHr39-a z>S5A^LH>dK2cA6Y9{Q4BO{PL?$vl*-#ReTeaFU%?HI)h)rMUb^(NDhBflopW zy|p}VUWuC4pvlH!FDvLiec%!U%h$p++wB5%T%lh#^o%J+T-BN*nSj1UAwGTQxHhyo zavI-jI$58*QoBt#6lTWr=Vog{Azegt(BEDlX`P1Nm7of4%alRz=;?+}SSQ8CfRC~1c*Vw`h;zT~$!fj;~ZxAB5<;BzS} zv${ru&Cv0h|A$7{bq&3FLDe*>0LE=!q=qGGGtkGg5j%a>()t3WPrBx^56(&C*N6dE z5EMG~<&QJB#=DTWA@ECo#MC4}h@CD4Vzp2 z^A4x-V2jr3H@|s#*zPdWIC>o~8fcOSe?UwxVlNew#ca6YIx(Z5M}C}sAD;sQ9cPyj zV*M6w?fC=x6P7N|*CPn$pcyDA?SX2TBXASrLHhhKvg-XgcFPFX7AjI7Oc=L(fL7IQ z+Q2{pQ1=3{=(EQ^7y1g-to&-(({q683FXCnG8SlE!6-BCyT|)iYnR2ig$zwx0Mr$Y zdE&4{;|9oISg1L=q&?-i`pRvzgpQzrciA_ITZH}rmCJY#(lm;FYRexEEa8&# zcV>T1D$-d9N5R@X*v7^`63=>&Un8d@=&x?WxBdZ%E%HbpfTqHIartES9$BHyOtXXw zbUgwyCPNhh>Rq1)ZU0D6F~G~}!SdQ8IrtSDI9VB2VEDr+F>>ufwI{4jY{CkZL>A~i zGk~<;H1CF<;IPyjr)btb(8jaG!3S7APRIg>rGids z@!i{`>Lcf3Hw@62`r?f+vn<6uUFYJKXu=eVCZOZ5mIZ5*^ynz|71cqfLhBiX2f`b( z=;|s9jnAYlq@Rxrr8G|#1qYuztVLIXXX~n_rL(m$vrurdoRTiLg&HqUDdWn*C7>M# z{@g!pI|(jiy@vmwDfsFy5IzQ<^9kB@DTO{Z&}WzjttK>+R20bbE`gdtdw|{*kA#BT z`vA%7?O{`=@J|MmGRXeEB&!+f@4`JpXaf~zs{@joM7V+h@Q0-i_wX6q=F7(L2JYH6 zEcYz{<+%;p+kK652Y2AE2iF;FZwL2O6m8&@@tZrZ&>!-5ZhZ`ap?N@1a7N;}#(a@N5P!UeUl0O>6N8!2Y&eB` zA8yvFsj!)>5o9mWSU zXul2yNQi9*xho6oVQ5~y14)*MB^A%jXS$0mG^hT+k)iiNxjDhGeQ6Lc?92jfI{a3j z?<@{}BD8vF$_SUN?xvE3MlwLd$ReTCrCc~zUgRdr{@8p%yf8yOQl)Q_;ySt?cji$ zuV>c5(Lr{Hymdu*gXJ0XjLV%fb6}{ry>Y^Que=O2Ue?`^`oPm;8Qtf*9w@KUD6X9W zd>YA9#r)(u1}n4T<80R-j6Q|iov42FV?4!_4|eE{!Tp#CR1Wot*)iO3c|Nb=k6`;# z6-p=`?bFlgEEhu4995lCHBc4Y`Kr&`{U>xT6EyaoAG4)SAo>!~wzvEe*w!`bl_%3w+=< zJ6ZifCsAV+PEm?uDEHWYgq5t^AEk^nkh>qQD^~D}Ujg4)wkGj%f8sdc_LJUxtX=zS zK~h%3P>RKh*oMKoVM{q+t640a0d1aspo7Z~ zG$OGjtQqJ>v8a;H5dWwwir@6Jy|p(oIaWcX!(MuX*5q|-!5KZ$Dc!u*p^q6eIQ;al zta{{91U2501%xy$Fa?#o7dpXAz7crGwwVnaeRgW^Y(b~#aM&~L zt9N%nkB!sI{URCeRh+c0%&o~n>stQyr17Dl4?W~rN4^LR(L%MgggvNrEdX{%lRy_I zBO`S`B!#^QGEg)L{1w%@<3Qh=Yr>Dz3X?uPE&)vVDbXF*7i_G)hQ2SVjG#L!)sqIv zzGT6sdtSFBBeZ8@C;onAaGRby1!lx*%8X5TTZ&kF@4LNkykL0*vc9M}rBfLyASz+{ zv-Bf3+JHsu<3f@FkdK@K2s?Y_FRAdC~q z3gQ$s(?%r6K*uRwtt8m%FcKg^Pdtd}9;(Hb-34^bR-WYpOI&hn64P+C?>zeXr=!yV zPVn-3`xQ5g_b9zDry}D=tNa`LjzUpXoZuLcd41!dn5`}ZvJE$pcClA`@5P4i;=@=J zI^csKza<0@XNGt9YTh~W)esBx89}CrJi#`@K1rL;1%J;g&W%doetuqgY`5|s(TJv= z-EPnOTAf4AOR}W7tE>PZx5r-jeUI3iYO55deYNvJ8r3U?{7s`t@#Hu(=*d>IJ0Ky`|?dJS5^3x(@RI?d@6iTN`KK>!$8 zP5G6t`~gWUSF~3|4e0*mczKd$=daj7OT@mS8%4->WGyxkl7ws)e@3PE+|ciP1N?{x z9aP4R76BhB9&6>kqs@Kx!!Uh1~h$Ax2r=-M~9Cs&}T0T>sKJ_)6^ZLDNA!0oU`=q&_bvqYcAMhVg%Q;JdmPLtf zNkya@q;gfr+4h29srt7=Qk!k=r(@sB2LPMwH-&P# zCAmH#5(}z95bL)Dscj5)Z0%kpmG;K+2aVRNy>BORo`C9_LkzTyx#kf$)FXgn<+4b* z@~mPyrfYPYolrl!#omh=XYH;pxXfpNeSb{I(s6C$dd}&~ozK(3t4_$c+MjhJbK)5! zJ2a=3kRKK>d#yhCoZzyM-;J`1%-j@&xtI<3hP2n!?rf>hs%hzEg*pQxCQybD$4Y2e zhnc-TW07km@1#NBq5QB7y~6OMrFX8jUwUMSFMnSbU9nSYTf_TpjmOg3Z5=M>yIhs{ z(ORa>0PVuQHLAra0?Yb?tC4$^6a3TFetEdS+6TknNAsFLIO-bv6pVYF;0NF7%2Ad?87I8^wlPeP<9Tb%(>6AcmomA&uKO7%BYy(xZ3< zl}}4J)4>KNFkS?x(PTA`{30qc&i)o(28ch4Y!VdqqyaYyquP#*dvZ5 ztZ3@iT^FNPe7$v%CS!v^Z?ikv;dpvPm$UXwTO`Sq_0SN=%CeD)zu39i#+$rAgw zB1-0^v&(o2t?PY;tRlBHk=1a~;4qnPe*OlNeKS{}a#u54)i+qn{8+jb7*LC0GeXOx#)t}nP8eA4d)t`4%HTR+yWv^{0Z*1m4$t@2SsqJIJ zYwx)ep_VJo)u&&-YPWaPgcREwBL1X@vF@XO;A!gEy8XX-oe%7RS( z8kgcBE`8~UDAgPhYjwEV?eTG zj1ME51%cw;->asU&#BORT)chY#e}$z0-w195jbyn+!t?Z+-u9h^Z} zY)Dm|gdU#KNcG`6>#g$9*{as#y5!OV8HsSRY6+I-6yC~yh=wxqAf0ipA3!}~Q5U&4 zbo&f$V47}S@l=A}NevzDQ|@@P)i+KT4zTp$Yz4`r>(epy2KWnS^n9m~Q!OXc!`p99 zOoH&AE(5^ja9GMeU}CqCwmqWN^S+@ubvWH4yFSvN^R*5AtEie5&&ScRLCL|24tTwf z4p4(fu_!?ScM*ciMnmC&AjefehA4o=;-~p*iN^$#4silUUhAIjOPBgkWuSZ6%F&8N zF)jwl=YDaG9(d|S9=}9D^DEcGyJpV{zMxk$vQA{D z3q?9Ku|hEY7$UVO?aU|*-MJm*S4E-uBnGgZr42%gK7m>#jaRwC8L+5^> z5&AHV$#*#_Y|X{NEIvmjblF+}qaA82w_VSIoyXx{RdFBkrSy#|sgE;P^I4`s(K}t5_3xa_pX3BFP28o!`&Yyf3 z4O9#o1O$1Rje(@kfAzud`w8gd5(LpNe)F&9v?$}let+ZRq-24Vj;S(Oo zs&Q$Erw2~iWXq}x*Zdlay0*bO*c%_lxQ_SVCqoE#s2g__V!Tis7NX}$Kq~1#ZVm<@>Z2A})0~M!)?06J zrV6K^U(xslq2a5-2r;NqzL^qX%!r8MY2$Ni3jKyKt5AeQp{PHNfh2G|ITua?$~V{T z#~^(aFVknpZ#w^i@lj*UMiJ-up{3}4_YajIAeOfgb%8T_7*(+e8|i+74o^!p-2&G} z_bYD&AO9hV(MLVDcS0gQ;>dm0UwJx#nfU`RMAu!eL=)Hi)g649zleK+&@LeGi={cp zqFQ1$*5nxMH7vaY@pX#Tygf~su43PAC$&*!SN#o;rss|7G_^|)`Pnu29xC4*%Me91 zBYz@}XI~+B;)mThMM@9^EUMvQI~4gM2?wkwV7NbuHE{9>+o;b0{VPQR{xr#{ToiqN z^YAlHc9iFn;7W=REPY@rf~NhNePlSZGfVZikYv@ICh}(Au53kLA&QQ4&MV+=Z%WPD z=l!(|PK2t=wojw3Q5`J3A;T&UKv)gRORcsncN!urlj9GX_Z%^`NF{?1!t)$xd^n1f z?!x6K6g5Zl>BEp}u$3tFG)5`On#-FECzdBI2bY;-mXJ=OO-48if z0%Z@e=`RYU-mH zK~b!hcYg%6=f2i4-|OJshR)F7ML_o=R00@kn~7ym!9JiTD`V=wcB@i^8a0Q&59dfM?Sov;2EnqUtgNa)9MNMuX|`Uts#cFrYq<;s~xA*6h;hhOMLY}b`L zUw`o`prRuUHdGaQm?n9uo{kVZ?_cdeU)v@d`cp20H2I+J;SY>=S= z)CgC|h?@(?uGF%+cHcCitQKI8uVq_>i)A}RtG`ZCVsMUf*_X(3h52@*g*}=a;Uw_P zeDa3}r`3z2x*V+k*bmL*HS;VkBF#F*@(WEg*1>eY)$UXXWquVc(fy9?X%>$Ch>a7Zm{Bv|1 zT1hzG+DXtygAFxe395aAzX3cbV4*-<>-v&;UPB|$;N@wmJDYM|$^a3A^dUiOwH4+Q}8dy@#a28qD#jk{hC zSXkudLo?rbBRTk$BuKun?i_r%YeQynoMU?QQg z)Zig=0vw`YFSVf0E%`T;y_N1Nrcw@Yj8_#XR)BzH!5x2Wq?M%Ld5~LYmUH3jER5tt zS6bNF7=M1mMk%MS5bAlt`hVG@O~a<3X%0u*dS^(#y?0LwR^BFgSbU|HJn(a_5afjS zML{B-=id~)3WpYLbMOjyQko{M10}nhWE`P^5#7rN7Q`x)sCIhNb`yET~w;Nf{H+E&i6zOfB z;l(Nlo1Kf&TKNylgq^CTs&Y?9za8ZTH*dGyOPOb2;7BRJhnb$x{=ghlZI6 zNhx{xZu=4!G)Fj zj9{t*vA8vMu_7M1eg*G{2eZiY!$L{^9(9;bKdY5DF&yq*Bg?Rjlf~;n9g6$u4$YA0 zLI#zQZ2$BR_ypEIN-qL%Q`ksckbSoj1u}V|uL;31&u%n(#&QZd680HE!4KNEe)L|T z%QW7&`@@@tw{E}Q62p5?M^1S-D8hSOe!bm*+&G`F5AlON#me7zztk5?D7e}k_~n8) zk8a-Z{W2(qR0ITrBKz&x`_tTyhJrB0??E<7lvKkJWz`{1 z2`2KDTbS)h(~Q_Mh2ZR?n$>}@Y~~%1`?eLeM^$C9aZ$LjkXyyJ1uL0RZ4w1LFYszy zifIAPyPE%l>EIiGW^3zR9Ig6OvYhbuy3Q$DkSLPSO&W3eO+ku;Q(tE4(XUH?9K~Ju zSGvn-kk*(fqn=t9nI4UD#1kHq4B@BgW%AHP-OK|A%}7!e?crc+D1wtyXlg;a*OmuV zX9XiTp%2YHaB?_A$@z7&faLvoK*&T6@A$Wl57{%XXj||l_PCHyNXV+2_SK!cUe1{u zyoL_J*51gZ$~~|Qd;kenf6!`UMsgeGF5i%G&vN2pAO994x9A&QHuRQ|-ipKkWJe5GymFDmh z>bg2mf#wA@9>$s$dYbHT`RsL_*Qrdm+Vz~;aU)aIzzbt_#YaQ`-6JK!JE9}gT+j3V zqHl`4zQ}txu!6Qf%U!Yd-Qmfg0b|Al)?3`O?5~hkN;cTMI3XDEK;%SuihGd^xj{3e zf*NI9x+4-WJs;Vg8Q{ECd_+0T_~cf#5xO9Z1CY#8B}A7j)X=@Yk=ojnCh}||P?h)! z$HG5e?DSx)zDQx?_CaeB*NKGG+KVyy6LTcrSmgwFBQ>A3dppU1iNCH@7Ml>up}d|s z(tdILIn2jTh-D^C#;b+lgOflVXG zxJ&w-G+8mNl6+c?zq{SsuHIh8L-`_TOjtu3wwYGKxPg>Ei6n#%#;Muf`zD+JZEf8O zgB;LJh2E+D8@<0&*G2Gy9!GD(0DUX9At)D+a2;r>V4?KWsJBGs?NWQ`1*992t424!rAuL!Hc%-+dReC zzR{oNk+`;q{e7jG?sQ1l#^s{<u670UGilt-4}ZddU|t%U zYY6lT@i`9;2IO8pKJ$GN0eeYKBWJX^&QbWV$?Pb5Uv%J1fV5~P6jd;sqK90sPh+-- z4*fKy%XlCX^ASG9V7>teP9tBv1fUaj6I^^8r@xGd&A-swTIWbLWtdDJ>hMg_@AxoX zs=FWEdVJfPBS@W1bW&+r2M{gqYi zuHpF_Vc-#7?{v0wPosg#6Or8QDQK+xZJeXI8+RW8hp^-J0C;#Gw7&9H6L@nD-xGBNErmaFV~Lk1V)W^@8(uLM;04|Xhi^L3qiOG@nEZ+ZXC7sGdiQy1Oo|79OS|=cX-O*v6 z6Rh{LtvyIlMU`gqT9DTjJ!T$pAx{xJ(?ew&{>&a|3@woSJOOPpnqii+6{=k*>5t;= zZilFD7Jt&%^0*h`61@>t2Q`@;d|O1Fo*lE;8#7)`;Cb=2ODn0&Oc1$+`>RpFNz^6GnZ)p(U~<5j7 z(4Ci!StUZ418_{ybu*Cki-pR`*7_9ih2-@dZv8ceUg{-jk*D%%lBp%)p%Xl@bP-Nh zo#am@bV>oQV;8*SQzZ7`;oQlEB&U_chfDZKU!<5=~n!;Oj@Z`0A~n74JIJV{2d~s6sKFH4)LavD z<*Dc>l>?Ngq|C%AunLy*LQ)Okj$Xn7GEl5i!2moB+}1wNzfnU7&FBuH|8XDS%tRGt zv}H5+xv=p-l0Cl3s`vSLOzK_iAmd|8^iMwtVEY}8YIh(SUNW}H0Bq(8n#%o$9KKgw zBP;EUwE3oJAb?p8nsx_kB5t~-DG#Vl{-LW;?RZ!g=TpC66)g%$hFbo0PXYBE7vy)? zPEpcO3rn|JZc&4C=-H94ry?Vdp;Aw}B3Yo?6BtjyWc4$jH+pz5a)kfRH|R;pNlHU_ zoHm?t?ymc-Y6n92X}6#HX#K)Kn8b0P7U zo3L4zj+EIOv{%v%N--(^u=9NBDB|KNO$9^}(kp{s zK@1-cH~-;)h*LoegP?-CW51!J_r474-pOB|ZR->8%JV_io7xDFGB@dzy}*3{FSmRU znoU`rHj?SI-#+q^js#e6SyRR#9s8B$Ng4 zYu|#0tTUD~0O{N}#O~g45BJMGv=n;y3RNBYn)&;&tg9gA4|BAew*h0dhZHBdemEDP zIH%Lz^60hWRR>?8YB@Q?o%$Uq`r=j=#HLMELZGG%^t1ca<{_91!1!)BwvK@@RuYbe;bXhGhKwcHhdK>`1ifT+2>nia%< zkf)(y5-;Osx1eqWJgl=CBQ#?Ru6|CZ)MNqNc(=M= zGyfl2Iiz~4<>(y2$8$|J4^);!mLlFEj7|^#-D}p3f-tNa8?vB9UeUq}wLl=Sn z^63y~x!EOWM6(Y<{T1D~iEIS5+0->^&)6Y691hv1hx~a<6q151W`l&?VUIxlbOXXB z{>oS}K9bZv0sBUFHlE+lF5v-RVOgxJPit#uatBpPG-eu3MPRdu+6ew1FHO_vn5y6J zR`T5ne$QNz+J2g4ScM(3hZz&rapztm5pcbfV>yMKqvn{@_ukZ(vBzTPiArT$92D?- z&rXBVI5a;|uOqpm0Q^j@v+pr{)z6FfJmx#(3mHe>OT|2N{B!wBl3IB1@g#@EbY`a) zbH98}CMqS}>7~q1`HTOHG{PIs8V$=Q+c@c9-yqbVA)p8TI^TDMvTZztCfV;;WH@@# zQxgpdACjwzCm4w}q3jE6@hP8Qk-bk*^Ii}+RbcrF7vzkfgz8fJE9FKm7W(gaMqDwH z7aNu|M0*9CQ4`33zr`RuAw|R1vQJ88Vm{^3{dIpO%`r=81~;Pyzpiw7t~COb+(Va^ zJq!DK!Wb;l@?b4rTVDM zLN&-2NL^0=1P>$W9(DCcSLd|?7}-X+@TpzhPdkFDT)1h;13|8PvbC2)BJ$W|Wk-E$ z^LJep)i-CM@T;WZP5EJU8(ao3fL_C6VtLq|Z2Oh52uFxCT?uZ(=-SZNcJ}$yhbz2L z6Y87x%p;F#qbpYwD9?kYSzw9*zW%+7D~ojB0kB!gKC;=x4MgyRVx-uCa>s&aq&dMs zm<|&{_>zyKP77b6+KlwhWh5Bo=EuC2?u)XU!9o)W)gc36_125wNrE2{MeM}BU|>WRaOzlaabx5U7?M7_Js$NjLPI4_zu3v6%% zdFQGN@APnlqatWerlpJa5pYzX@s#qfeRXDS^W2m~Y(zEoJlRoyaOJn%U;3u+C@1DCW zUJcG@vae-en4pmGBK5Av*W7MBXLwjSR?xnB3AH73Mm{{ zmU+FCFKxp{Nk7tF%aExGNk;()`slH5_aLwnx7{6^p|!IHuV<1UAkYOQ4TNha){7Mh zVf=6~Ms1Cv45{}mRmZ}X=RIlx9gtjS5m&B7c@XjO@(&pqP3yyhV_@6lW|xB&A6+!| zZ1E$y3=(Ljq~P?RED$^4WzxJ2d+8Dbxh;FsC}QCET8}-X0{|(&-*&{{-{cmAZaFlr z&vp50hwgK_s4?b#44Qrazn^Yx`R-jWtf7p$CJIYGLxqo8B`<=Enq|8!|7WkUj>&vMt%u&9N)OACs;gLMuU(R?^6QP_B4f8dVkLk zQHBl&|42#@;VT`u64l#;z`W$|`bOH8?ZvXuw09m^szQRe7o@9y!^!2k$fs8`vpt|8O}w-cu&F=~zkdZ6T>0%QyY$vS{9OPp&4QUKHdgVWln zLj1+N$8V&T461Vk={|r%>GXwBGKcl`BlC{b zlKPc_er#P{`(P+t%6&e;pR21ty?#?WIV~9K_GeKrpkoLSj&md$(8UmyFyeuiT+PW~5rd*cp&#th4NICk3gtc>DB-iIlG{^oRA#EsY+B z*XjKfms)dK4V#c3u4ojK@npj5$CCM1Oz28ng82H-wBGt+4Jt`F-=9%9qi&QsB!70b z7%3M^;=)E4Ay9w{57o}-0>FD@zGEJuU6fWoaO8Go7Nn^8_UVfq*$iX{x-j4BI)^tE zKf)jO&#+}i#5Gbg<-=ufI)D!A)BY$*%t5|86{t@K32{FqdhO{l_5i+vUo@{X@ISqe zC5Z&Q{D3Nh5bvle^rpIv{`E#eVCXY^uKbJO^>ZCE2`3~W{_DwpV~Cjx`GgoJ7qOD) zU@d>ZiG4Ly%p=O$aD}g`#wR5x!}w-$vk2}|(7;U~dc6<0@Z9G-wZi{SPeVg$@v8tz z2J}|)gwY9F_m*6{-_m@?djFjEZ!+ocF}*x#OVj=5abSbRm3_zbkmhi?%V!u0un%qx z1PdS0)(<}Mynh^IHQ?pV0p&bYo2k+*N9y7M8xoV$Q!WY@nvyh8(fnhjJjRpNJM-)4J4gva`4?TFqu8hRKFxc-;>5osEz$X8qQ)9%7Mh?3^u zWmkWD1KKwaaS=#}*F)krBpFJ2I-M!>HJ3NiV{f@|yIzI8tk|DD`hB)8uNCQsf>sRm z<)WWksUg!71|$esgtPHHhrd`Pz4K{}LOn&=%rViN1U2yNI$ht}epQ4jmfpFObMRO*>}pl6U$QXh{Sk>n@9U6i4Nu zY{&f+^kuaZ_|Y9&J82mKFD;Dwr(?9=qcwY|pbRp?gYN=G+H=rJ@9UJ< z*L?E;!JA>tCqMOh4m;S0z_{I9!qBa&S!X>ix_(?_3LAT z+&tYM!Ai~c_11NIWcy@>1nE0KY4CuvH=K3uY0jFvQ*z5kVq0-hF zVwJg5dIYy6#9vZph+o5=EV)at#LHf7!g!n9j(@xF*(K1Y|E{+=y`RJVD??8QFF*NDN2Rk+g zqk%Eo-!lZ7qzBRyNRW7&PsMZ`@c@nU-F!^XJ$x4IB=I9LLtQMjf1ctX;XJ~V*=@gL zzInSSf25lbbUPZ^jajtA=kY8DWvHMwI&J%sCo+>`t{xrfX0 zxmI;u7)q5d`+Sj-#R4tH$j0%L2nj4*K@v?=)~N3+2|eUev^ovB}Z_#CK`qhu$x}cfjWF_#E6fH z@YrjPu1!_b==LLp&pc;{3aGYB5HOkJGbRkU0qeD+_Q2C7{WkQFDo1`@0%lYWM!eK# zb@_m08y2AKQB zVUews43__Uk~i{(V2=InA~+G-oF~Z7jRA4zVK--?Ah?4 zW%&+{Ppe2Rv(-BA-d`4~uYM-p?|~MB6~KLjSz#}5ri*i4dBBdg8VV}c--rgqi@5tA`l z%+BZ|uU1~v>!+4%-Vb0c4c(yB95jV%5xsR#^v_SU-jG^=ZvGvjZ~9e%w`3TJmbHs@ z%opYfh=ZM8B{(j+2$S9Dt-;%=hUC~5<3MR6rfj@C%ax{zAa49~Sfuy!n}*!7JcBx& z57{!}GoIamL*=Z!8&K6&vSq%56LK%Z#Dx4*X`hV-dUkZUT~Q{G1bTc(QH=#9J&P6W zf~ax2EMEyZx^Ksi?IIpw_2|`nzaFaVm7LeccdQ}^C}$m{xm!o*58yXy6)z|G@^?gh zpF-7WH%le^zYZh=#G(m=6=Hyvogo-Q?t2& zL)&Im1}EssN*Pqicih3QW5wUTx`&3hZgBM4=jqKMPYj?_;T4feuG@E7Cealbl6ZkO zuSJMBk@M(5RST+Ota&etqlp=w8z_MaoG$X-I&jd_hT3Vd;O%HwKbLFz! zSXUM|%H5nrN!6Zh51gLGzlkSVN6FdbF-9G+Mddey#+9!%>HfBXm8!w@ zrm6i*C;N7^o}vik3&czYbHZuwn5l4BtUzuJxw^4m9!}Yy(gdgIcZp^L{EGr-!c|4_ zUT-Dj$Ad4iAN+4cP1y^|sT-o4xp zUmR}GC~m6@o~r)2gaxW`r;6P2>QfgOf^hG0hAj>>5Tn7TP*0Ab_fT`T-Vq}?a56uT z*EXf)*_o$*RM4+rr#K2GWG=peyt3Imn}_{G&xjkAhOO2iFDuBK+<`}oTm_4L9LERp z8Zrp{-45L5yipL6dga<-cTe%faCna{o(^~RTb~TW)D!BpZU{GPhkaK!qvAO{%)SHH z2M`79b7tYOd{Ip{1Ly9rw?}jZWdu}6%dbyYHO+%5fc}sw507YJ^Z938pVUH35iW0! z{B)AWomR<5q~>ztQTm!_7--3bjy@G7b}+j`8ZhR9ktfHv)=_jL^;7lN_M*NJa=G5} zQCo`z!CaoN5IL-rJn3^?Kd*;hGhy zX-Wb93lX?X5ZLIwDv3|l>CX(A?AH${a8!u`QNTwI08%Vns@Vg&tRJg4a+{fma6&p< zm|_>guhyD8PLGqDpzp3o2Lt@rExCCeG(MGq?fsENRtB}E>8kZjM^4$Z_#_MJMOq+Y z-AXRo%cmm0KfTPzxH^X$8rlKUw_HiZuqd-UyeYgw+{?>3D&_mr%`6}2P9CHPC@X7u zEng&o=Zj@grsRPVB0@V1~;j#Z}LR<@9hJG(1(RF)y zd||^tvVz)?wS#m!XqNpV*1ZR8tf1k9B|64iA|At;yi8nCO5O0BpD#HM+jy9nh$6sL zm(t3`bdXeZ_B|*zJs|T}&%Z)X-$z-OH1?$#K19zmGal_XwG> z${Sfpwe`e~UaHQ0Z6?#KRGV&4!sIN&a3icd`6YTOcs)`;O+FxmVrs}`%DSJwmkA(Jo|z; zlAioL;USYKMLIoXlW;}^5C!0rUaAAaxvFDukm|$C{oGeA?46f|Ew-vhA0ICsy)Sq8 zDE#$fEXX_ZhFdXcr{5bGSxA+!hQZbv$(MQ24fiAPF4HQ6X*nS+C2QQrIkcl;+(2xc z=h^q==TEcETY-X!nfjyiP?VuFAqasEmN<>gJZ8%C#Q?)q!M*())iM|e4uC04gl?XW zN9&RixnkgG0uC-}z1S?v7G>Kp9W#Q$*nt!K`w;wRBfY`jOEwF)eZd7?4O`}p0Mh9gAWKli`zdt_P)kR zss5OO?r{AJbw(+wNDuzoE4`dLf^JV-y|+bQ^X3D$LhGs(IVgF5YZ@{^Tb3fIYeDMBYNCskI==atU*IE0$Oiu58ut-y5J|8~CjX2H>zODN3fq6sVI?>c>fi7c* zc*9x`R6=2QJW7{w#&%4j8q3iU>lCU`V>DpcuG9I6he->a+BrIvc<7mRx2wLSij^xK zpD~tKA-b=*qHTKU_bgKGQBQKpbQTJOUiDW=3_-WHv5F%l_j186 zAWjFdkTV98H@C%zmj9p9up$F(|7b#R4j$(FGzDjK93O;4H?;bboby{Gp6;8yqu*e?##Z{RP-G;`mef^#Dm#FT^O#{kbpd z5<$C>Cs`j$hFCS6cX>-+}3)zYlq05jPJ5c-uC< zXabF!g(Ga5)O?22=h|tfIeH@n=yJ!JQ1_?l6L-)B!amqzWCD?a98ik9DfM-IUc;wwt*Przo*sXZr6>)T`ygcLGJT3K+Yh>m zee)8dBzCKdSu)uQ3T2aw9{Te1^MdI|D> z(XWXSyM4oibJh~ueNbwYpT0>pm(VSUCMjDGh@ccAzAhWgDp79I@Awdl;K<4s)-h=M zygVF=T30s36CXu&6N`OmJiUje&ec?X@b z1f~nGOeXrtV%^1yE$SRnWAyMx9BsHG5J~Ita1^JH($At4ql7oY*$-#;XKMtL3vQM- zKE!i)a*rHMQWa_L>0jIsqDI^r061Y;f3N#R$S}qt@FZ1(6qON#e^m@nUxKdco)fFQ zb0QzT_8#0y2oeQT`8e3I$#vR&}UIZJSku&8F<1wahM?D27LVmR7s_yJPovzHr%(kqsrOOL2=zLgQQ>a z#@wr9fJ7aiM3G8H2OZ!|y#5Fd>ULqeKQ0+_uN-E9{Hcc9Ugx+9Js}`1k%#hIA4U4= zg|Og-_f7nY0>a`@30+;#8nC8&hj9sCW&_axGRBc#JAhb2tY8-Q5KM6_@EucDkxra5 z)ywQ}&+~SxvLB3+Mo+-Z>oO%|WfUZ)RH~`@*NKKdRgG1^E?V&)!z=)@qJGS*$e-fW zrz4fKwN=$Z-Gc?lB@R@X_l^gbkHiJ+F4*j6g<*^ttd=eGO67=*9^FM|T~8|a1xX=2 zLOp5T)?Yu%#O(F4t0*hR!gcDetryJa><4neh=dTR`yEIo!2!DZ)B^xxkJ&2_J?($N zNf6rURILikaawJ5>fnH+<{a}kIU2QdDqxDC_pmopLlqnR?JX8!gZ*5h+3#pCRi4Dp z$xUDe@ER*;uY4n-z@W#{URl8(mVcBtke$yn)HWkt*)sbY?tJAVc1R+Nh0C$@V&j9J z%iZDBfXk9aG(oon!MZ^L$)umyLU5?OC`@3SgHe+mN%O@BM&QZrO=ON7GDH>pcOVDEiT2zhkdp0ZXpFfZA}9g_K)7!DfW2(PGKFTb^G)@^$;Zv zdoOck!di7h_(;Rw##QU-r_I;`p!fbkmqiWq&;&aK%Z@!tC5;7f^5TAr3H6q=PBaR|>oq(y>c;s!(b^b>(x( zKjVar#rRWylj3zUKwuw3%ge1*1-{HDhdInHs5F9;j0rk zs3;R9B^Z_aX2B7QLb&X!DcG3OwS9g-vER>fIVFz<(5*&`L7qohv_U{cjO3Z7He+LG+hE!Qqj41_1%#4N#dU0Rg|hgX*<<-P>J* zWhtDKCo^J)Mlr9>TmVTK&0`pJ%NZcsCHU1g^$=*y)Sb}Y?o0%UDJXnB@Y+b&-N<+Z4cCty+1q=zfWwtjW3X$`a*rGM9*GJPKMcp@^Z zqU{E>lnbKgjp1LBE?9TY&cc1yRZGygVMM z$aPn72dj%Z*TZ+LTd+jnHGR*=UY6g;s{};2pE+;oDF!{z+E|zb2$wsM!A;G;b;qR@ z$ks@DAxIY5a>cao=Uosc>ToPPa7W{#Fa_;*IPfyafbK#+Jb3jrwDEQ?rkspnU%#=3V%fI{5c7U#sozbAHK}9CjjPi zAC^OMr2Z`JN7x<71Q>60ag`tAg({9&<)3CSkUorKYCp5yj>Obo4UcP2!dWFP?t~?U zh5cAg1lHaP!!3-7b8#p+SL6ppFThrkKX=z7EEu)o+#omSx%jb65=~L({j{8u#-Zo0 zM*Kj3U%9N-;B&S=8ZaeaF8QZ4>z|MCvSt0;vec#^+}Ol5ETAjHqXQihCh{9^2^&t_ zcswD?I%_#G5(>=IT`@ldpP&oBXX0q_4!%kRr9f}|OvxuCu{-W5!A^4ie&ywx2_J~G zG<0ItC!&%}dPL&quVurn%C#pDKArdj_0Y7QSl_K_Gole0);s5D>Leb-OUzd^#K_mk z$+%JH8$HJ9oeBVJ9fw}rONzGr40?USwUEXZ&s0kvWed|`dWm~*p)Wc)k`VQ{!Y98T zB(Nel4#cJ{*}lY5c;~(UjeU~&Q7FDxA$rD$`wBr{7KTKbru=-lm%&{qW?sA2IJs9D zyGS4sARdJ3k(r}8_8oEup@2*Y50J)ing9WmyK#QNzfmE*U_;e>2bm0T20yzsy);Qp zND2}a2)c0k{T&rS98QZ|iZ!qo?&oaQ1)}#qFY&RS%r(h{IL)2c&#PA7zc;)|m~R~^ z+VlnXI-XFahXuQ7dtkg23|TDdS>9EO!IZS+-0s^+J)SB)v+6 z(0D$^8xiL8W3C&d%WXomTT%kD4`B&YY9e0FWFa0{h`EWBQ#8hXlhI+A^c<+AP`o(?T`qR~ATg(Rw2VGDEw!;gr zWfb#<_ETb1xFuJGLf$65$bp-VUpK z|I1NJiNh6IyfuX*FKd(*!UYhr&oyk%Qga@`NWU=foW`T7DYDx!)q&~ifg!OG9FHM# zk(;m1UgdpCQ53Tyvkk1=&_)PW?fOy%7Ft5TE{EmZ>C*Ke3y%hkPuK2?S_GqTmGypl z&;E^Aw3JZYB3s5Ag z{rtf+?xvOZu5@;e>rDHW4`xu1G8)|5%6r6lY;hELZ0a2ne*R8r0m{A)7)F3I*+H0m2$gQgAoHVMB6BRPIJFQhZJ9 zM6|AM{3Hj->bRo?SifzwBq6ha?DBK=%v-ZB&@Pq8k7X8(l`M6U)zTI5%t&FlKadkL zF6T_J_%EAw>82KzjKm3rD2SgF59S3cU@XS-&7aj&dLK7%pB*Npgc)9rb%m;W$kzVh@|f`c${)H1Y$ph_Px_SjlHcxX3odKA;T?1N z_JtI5pLexBdDV^$E+{P^?7~v66HLX*8DW_A5O+U+w+o0Q%g_mqPP6s9UrK*HWM`)F zms)d#9p_ilZHkU1cvxj|z!-Oaa5zjrTNIlU4trF=nULXK5mPP=&Vb;jDaQ=iy~K*K zo&h|*k!THxCG;{^uKl}WTHb6ClKOmZf#59ub6 zPojdPXPc(-G~_?EVH!+=e{^u}kIZGST;1zR;eo{MhCx3qVPA*3b6O676;pYvpV{La*R`azQKr_ERX?M1}ybHTME9xR7$u`i^Q{+Qm$rEOL0u0rnm}9* zXA+7qzfM0cSEu|(YYhK~EIIETG`xnMd&lH^WNxozC0*b%%ZIk}nYFz`wGIV`7@4M5 zm(v{1@{3|7eAw8{=O)(XsdlZ)qq7h?zgc1>04t17AN=wS7AauXkE)mg>6^&~T0qoT z7v0z>qGE!gz@pbL>iZ^V0+(hxh$zhi;dxyi%-Ns|(=|6O7;Mb!E?eb&P`JbF#D2J- zQI82RIM8hoVoDy)BCjDb!9JI-^_e@<1HdkBOi?wJbR^zL95`;7h0r|5d2xiadlMIT z7h2j}2}paC!MATCP!DnN0;ujE^zQQ`BfmgmwC~YzzCPH&*Yk5q8yJs1Vg#c}BnQ6c z%>3b62q=+lz!Y~Y4rBTTWukUu7q}dp?S357Z`pyDhYwlB(~B6}hnBk4XtZ!a5`36^ zI(7FA0SH5}LLyxqf4%awZ%-Bt6R61#!_~_{g!%Y~&BsZ7hFM_&jm{?Li-j5#oo|`@ zcpug?RTqarY?p2ibABIM&jw>SJ}Pi(eo9tawLK_|M)z|u0y!@TfZw6WY=it zRCycs1lSk>T8O+dIvl{krSafO!A#dfoRMZD2M53}xaz8 z%8#$dpmQncO*Qin5fI*^6QSOZ+5KK2L1#kn`=J_LOeqnkXk#Q!y7geDXOAEg1fERk z);b)EMhu2p$83E;{E%fc*AFAbK!|k{BW^3TU_%xP8Rf7?wf5CzPPP6*IIR`m_q%Kh zv_U=qTFey7PuIBqM5<98jmPrmx(y`0B691))Yb#8wm?)et&`O8NBiLY@$u(_j-d2G z7X!ji4b@&P*ia@p{+;=IQ1+}PVb+lMfZj1RMl{pswcil!K2?y`Nr!Tj=GVEa?`T=X zwOXtNu>UWKEAVPnX^Zx-80XIV>0gKu^dTtw;c;00_WNGheI+t3vn&*jujWbsVQpH@ zvwOH2Bok*xq^b+K(~YeFXW;p6UJlfA(P9>;!)LJECAZC}5d3@Qd8^3IW}RtO-w*U2FCO)(??A z)rhpgGttU+*#u(V!K6qfyw@GHGxL)a#NP}z7>5XlWI1L%6vOacq&FKl>z`~hzLdjm za#B)Qr!|4jKFxl|b`tJvKtjLmQ+u}ei|Cfq=g!hPKSR0bSsxW9#8X2u&*XtLQ%kMv z7n!$~%dnL<9H9PRsDP%NP3JDBlEq7kmrr@h5_-)ub$v<85|h7eQzPB z1g&14M_Rwdxq^a9jL!&)KYkpT4PqUk(qalc0m!*@0bvK-N|Sd`G&n+2(N>$Yxr`1~ z-NvF~t}w{)nc-48L+3scAJ~mjg#oz+>h=jH&{He{%V~$m8T_^|sKDfWEZtTAZIprf zF);)Z<@>KfOAO-+yE1NQBgG!`htXbCyFMVt_#y>b&f487`feXW;87fHtK*tae$O!a zwc)NDG%+G{n|5(;R?$&@-Z2?TusGZs&yHzIv;{wJ0)g~IStCa@j*X=LL35k6DDtkK zR1iwqBlFM47Npa@2zp4A44DF#wzwa+4*GlVG#sRGgIauOP-O+6DVs7+Ky5RuE<){& z_hz@v!IW}?LKI_kq*;#)m>0N})Ol|~quty581mOIe%<%3QLdx~hobm6XHS2BCmOW- zt^qtv!|8*G=Aq&g24qR_4s18=(R1!D8~6?(yV-rayu`0qc~E4x`^&402Q9E3m%V_= zF0bDXOzvnPB-a^ARF8(@Q~-AT^IIJ1h~~IPfj_<;y|>A$2)o-8f=v$35rN|r>_vlb zygw!BsZNW_lK3rEXX$Z#_+9}2ZGt`o&L5ZxLJf)JEo+EXF=MfTvspd=3Iqk~$HlQK zhdOf=!Vjj+J!m0x_6??8p?C9`8x_(Eho04LS78wu%h*3@HY}lsE19Yb`UWlH&5hg`LmUz#s#nq&9=GqaERVNLUho1asZ=A{4u2~IGeVIa z%V_W6{uVJa#Gy-KThlHRCVe2OmA#>{c{TQ##?M>XJ1=hqGI8QLNQlmmx*LaHvs>Y$ z@yl`_x~>`JJ^S5vKMjwFKgpJ&UyT*_A#w>O;(Gg+%{4;-=%KTC1_26|BCtNaTcX0?(=Xw-GB+r&A~soR0FmbG-w9D;hc!v26o$0B%V=-g#O(T zBdEsDA19!A6tWtZ(DBp%R(7@Z2UGNdsS(jb0}OTt6d2m45MbVB%0-MK?25p)koe(2 z+hz_1L^IX$fp7ngo>6^J^|FNgRID22aNzq=peiYXMAZf=#@k*ImVlS0-6WMQFr2X$ zxP#|RbU-M39YOLU38eh0AG9sr)MVTTQVO{u{`!|6N$?2vNe^#Z zdoF`4Al_ba_%4MO(K#m%6!D@sPz{O%@paEj@m(3IUe~BM+@km1-|CYfeV8P!_c<_a zX!9YE5^T30dhpQ@?K8Z-H_`Qxr7fr^BQhXjfJ#mtU9vy4abAvj8oands1)dU@Jr7q1*_mC=S1J>55k0P^ z_JQeI!$`zU5oUcD&!@-7AN@WS*`}*BI8-n-;F~?uR{Gu{=a+Kq8Ddkb#J(5{lSg>9 z5}`>uyjm};@@F&4`{L#wZja0J{gB~tqG}DNHy!+w;DLN-1VaSgN0Tt49PZbvv#(je zfp@|19wnptYxUex9bJc7TVQ5JR@Hgt7I?+u%nZ8UKpXC5Kv%jZz*-t4jq$K!* zbZwpGEvR9K{M~lnRe4l*W92G!+=sY-Y_^p;=0k|+5f*+Ndr-vSK_U$vf3H=t@;g3T zWbrA_R8}oehAP3vSQW>=*7%Vvthh}Oltc4DMr(E0wJQ5tx{Ptzb z{cL3Zn$Oq$Jl>ZTzz=-jenV~#tNC}JW*}BVGKoZ#M3@p!Tj?BJu2d8m{H~hL83n3km5C0 zQn{qMTk|1)y|Z3guL@-ZHp|Ki5WXDrj2{QHn24-(wdOo(w~UJPrNG3QDLRD^IjHH8 zGJ49Q97LOfjTdaIwWEfoc`*|PBqslD(7(O;W?W={MTpFg_xaqx?)Jh_VZQr=ksQo; z_Fw(wM&%Yb0Ke(|#G5DX@0<$x2NbIKPlm^sU*e922EmNIUHD>aU(kBiEuuR2SM=0^ zI0bl29ii=EZhCtJdu7|xu3cce)&=-U8}X&|CcOAIY7T$^CFOgWOgr3L-i=bhTI$U% z^S)+Qekr5u=&g9%q^MPgRuAc3;mfoj(3TDcf?4-2l7G+NXFeJm$z{tU19T9-9kORB zZ9nI*Gx>?}KQRL2SmtuD18w1YHlC-~=L9fL;Cl49WOGw<8koPq=`nDoo)~FAJM6)R z6N~$Rs%IURS95*cfxRh~!ip=eOy3vVP*hwXNlf{_tJNsNz?eAnXpYM;hUgflMg?{!?l z(NSGa;D_IQ(Sz1ucxHT*C4w(%?3ZWLAiSyX?U)u1Ob`HT(H4&bybtny6;7O)o*U)k zDS0@g^UAmGOpS-S^w3jkwax)M;Wyv!uUz5tAFz6)x>SLFKzc(}wI0K?m8~V>=YdZ1 z(fRfCGi?w~u(}=`vuxP@_^%wt6PZMJFFl)H|wBHf0{EOa~ zlQw?~xdfl+Z|Eo9?GK1I49J1ff?GUtR)X+D+&BmAkP7*gF^?dfId`WP+zfjX2eon) z_OBdR$E$n1-)$!BmkZqDd(3y=#5A%xj;~BFRwf)BLqYWwZ6L}6+%+hhz&<%JMsJ4naA#|{aaY86}F^5f5Ho8Ot4Q;Z~@z3mq_wUUm9DQtn!oKYb5^}O zb^*@iw!2*qJG_*YSi~!b8ggxsL^j^`^Fetv4ZojXnz9JU9$`(KEVJVC?SAhE!XxU^ zN~)iAPp;$bUj3LH@tgx^!*A8YWSso^AKD)eu)((sa_r?ro3V+yh-Jl#C*J#Xp6-m4 z>)!MGx8ahefb1JBrCBMs%8bd`rW+F-+viCPQs;|bd)k8ZxWF8FtvOCcn`>p>!E~UZ z`fG*LOIaZR5?d=n;Y>Rns*ZeG{uTR!j@xl zg{VnS;O`5cPdTkM)m0AnZBfPHzlkPjGmz0>903ymgc}^;3<;? zLR^5~MWo8mA1&P0#Ubyw$MZO^dCAzwg=e2C`katS%FOgEFEEn6#nrd+p*LtR1?PkJ zYyFDuWCy87y33C+JPX4JfdP%;nHYpYmSviMnNDn;XZu1po-;wj!p-9E9oSL`63Ec& zOz9d3ldF4<&)o4yvKEjw&?irkTdA<=5&~N=I7|J?jwUT6qU@Xef=-H8dH67PJZK_! z*)32X(@$Lnm+KXPUhokQd7kRAlVXH9|0KBsguhhAcHFLaM?8$70`l1PIlPx0j-1^M=$jcy8*q|-KSd)kKpbVz;l2pZ{vvBgK(Y~#M<@1F36R4Q zfzT+mi@1uc%DMWjaS*f7yeCS&T`xcG$ivgA`Kw2Jh78touIhV=O|gri87x(Neu?Py z2Y5kJNPIXK`j-FXAkxNa;b31(VgNljpyfY=zPtOLU96FP<4F5Ux%AEc~F)^w{G%5Zd$6 z*KQ*)Kt($mqhYATVirIdBbahw{8dAFX1`81vnl5QS;#`GMtKGeh=OyRp3Vo6!_Pl% z>LUYfx%vqQa0?Bn|4pv*!QpzwC=&KE3#K7flKgrkUOI>H_C<=QJn2B6+>b`A@Dn;3 zhn{{t-nOxQk%iSKOw=db;*g_e?|}`ANu_=nTE}WzF(t0um#lqChKFA1^a=Ozk$1**x3g-Zor z^}vPkDtJI>_1D)spy!`O-O=H3u;#AzsR?gAYBC_Ft3xgkRKjo^Q^UB#0HH`qJWs+z ze=mx2M<3$2`dq6A`^`1+F9Ui&Zj3Kxy$*BB_1K@<=eANUgBVp=SNS*cR2bwDqMl`2 zg6Z!^sglr(W1Bp$ruu27B*J3%jgt9(_QDWv`;!vz>|LBNrE6e8?s3^0391WT!QMy0 z{*A(N$1k${2eax=90cn#6wcvgdhX6l*g*+>Txio?>iHl2CX1lphs*N2?gd5e4p2^M-9JL z47~xVnIBBEN=V)gy=R;c^wBEeMA$Dm1URNM>aNA+FtAd2@_3O=V`VIDKgPgm)nW@_ z*MrWk{b@; zL`b0>N(|wZA1$fzZ-m9!DbA1;o-DnCE6{*@2zi{|ax|YevA{aVzV_vn3eXNyZ-?~v zipl!#6d~plMxe0 zd1{6+lu_V8kA+d)%dEjlr|Y|k`FlKBVKqZsi;2)`TwH-tUu!x^F(24>=6 zJryt3Sc*P~!oY_k^W7i?BIPgz6}BE|NF=YB;;!YzgzR12mw9s#lgohLL1#mWC{*A& z9=vD{)tVH|$G$s9I%}k>UTcji=6L4Fbca$j+(>5S>Toz4@nU277`0g(MCuJk|LD68g0#jDaJODgW&!IymodUcZzoO&Wx^j z`?i*|aOv~0w6N|aP5a&j7FEBOfONQe zxP%c8VV&eq(9(-*=;h(Hu1iuL9^Qt};b%NOPF*ZmtOgOS9a-{Gd4F|5s~KSI^e3}0mtTH z{AtevT+R;EAQ&3G2xyo!Oc|Y8Jm@J_0C9*ZAephtaAL(Z(AXl>B5PA*yOQAK2{lS- zL=oT%?}M;%`-=z@;c&-j4f@Z&e+_ZeC>Ff7dpqXu{+Da52DuOaY+hKVS>w&`6Lj`( z;&Wd@(Em@)ab+=G6gOc1 zJpA8qAEcO5+%F)Kd9^;@S?*j0XXE>yMw94VJxIBHC0{DtZ18Ut*+MelsUN$x-LI>G zOcekIm%Pfw!!tt|_z|(FcQDxU+(CIL#y&gPns}7$TLQyom_7ugJ`Y}IAZ;8&Px2&O zHbiVnEj-@_&WR+R>X=8qpdx0q|9L=<03jx`Z@QpN7Gf`c`wV&ThHTyXR3gBGZ+Gsc z1zJ?d?p#DyLy=F;WPbTQAET^h`dx-7Q280G+rWPN=@T1qLEp`|aXtV&*r+jByuIaI zAF~mbD1@rwDTK}E1$q@mCLyBXut?&|#w5+;H&5onuEY_K$phfDlUX3!Htt>+?vAVc z-CvcAy0ggDy|o7i11c3Q6AGC>NV{yMynF0#v7%J=s%9VWYrXf7FMXnq(iNaeaQrQfM~905Q* z4FsMFb^aNT12J6krLT87Ka;(>lX?%?R{X=RC&Q>kzw`67IY-a@gVmgH zec%gzIzJs0B7T| zmNlwLlOkqAD6zgd$1oXYOmlBf#-r0&haFi%BGt#+( zUi5tbIG@+fOcDLwqt6KW>~$~Mi_8aVvNj^9=sRP=nxc)s1&kDTT_T+Sa>&TZPqRu3 zPKj?mNU5PN2)@9(u7vEPzWcE)=~-Q#h}ZOLWcC&d_ECI8TE|#}7GiBJ9t_3w!y}h1?sSpd>I;|s{GKjbm z-w*$E@YXIWDu-jLw^{3gG(AM;@4NQ9XwFI?uHd%ISkt;NdJH6t>-fkw5HCKC+bjU} zgsJM*gT8s^=1)KR2*#|a8pIgk%_{Q#W? z#9$yIia{&^9#*WgR+#P%ro&zzCmQY-uz{cEqmRDYR2HwGP&g2UKwQCKrhzrWMhf2P zqb)=gGL&YYER_@Y;Ds zH>ihylVSow#|&!}|0&&S?-lE``-B?U`{bEJsH|M*?_Z92o>migEaUY5~Vk z@AiG9DEgrJXMXE-7sp{u*yslc_Z-|w#jJo(s)n~0O=i@{}TcmQ4zz69G~JNQX0Qnx_LdL-5dXM2r}{eN4>5O z&y(tXI<8a9C$pE6y!d5@Qt(hX;a}W#u}g#590w%#8%Wn$J(_#*G^vD*H59pxktXYZ zh+B4H!LIvlUE~26Z%lU@?qLSe>92gcjvQSUmzOT-HW6Sf@z4oi^REp{B+TJvU~6L) zH1u(15cruq#ALR&;FL&c_bp>RtyodnWz&}}E?1=OY1DW(9vTwF%DY#!0d1kb_%WD^ zUu@M2i}O%3l1=TR}N;I2yqWtnZq>fMs`S9{-s#*6PtiXD$1s=))9l zx}RtqjYGnEvE>k-;ossp{HT|MkvyC)K7icg@D@7^2x8?kz75CfnZ~B~^fbHFpODBJ zAnm>Iw9y^CCe*LxF!L9oz7BfLk6AS2@E|^Nxk4&~xJU~GPN_PUgUGC5hlT9@ea0XX z_$sF%2;|T8Q_+}C|ASy1F+9%n)z#P5T$cwHI7t^yf%_v$Gd2x>u*S-dcKs6umAuZhg zNJv9T82uo=Qyd+{VWwf!ibyNaK)b#Jzb^X$$mg4T-d`ZnQ`r1EViLo;qPW=SHdy!m z2sH+Jec4?0{B)JLkZd2z724lxRA2yH{9qjf`zl=4Xi}j78$TW$BFTPtIIJ-(9~i5+ z75#YEw{S)CYFC4h^iv8@%P7{_Hr?b8~MB!D7T;{<_zsT`yq) zNluT1z$&UWHk}a2b@&N~*{3DD$zq$H)fn?=jHp3k7ST5zBZ{~3Rize#JwVx+n5xNJNPFkEU#sN??LL#C68e4d{Ws_IIL*XFGtDlH&; z=>#xWWVFi(7Vfc6?Wkg(f!uB%iW(#`b}T5Icn|_6t&|t8)MN=_O=N1w)OEcfa5y>}^}Tte(%G?N2c3 zR)|wo|C(z+`|1hB-Th-DrtuuUsrO2zxG#_f{Hz7dXG>ciq~qm8h!5Pv&`&{Go7v$p z7Bu9GL<0(x{jIZZx2n+*rio}*4vEOm;Xq>a6w{3!`8F~5k6Ps_poUdWj-H##06x!7J8NE zX)+wJvXQ{(PA?l0yzYwQNe6X2-1`fQ*$A-BbnvD}@>)I5)SJSJOK)sJQ-x=naO|k*ZWEeg0N*h)7 z7u*LBdgmX3Y^ivw0!gjh@we>U59>Mt2p{ejtFJ;SQE$(KB)|vx(EuJ+XAWdX$Yw6s zt#~*hNRbt86K7pI<_x*_?IlVKYZj6&8w$?Sw#P|>CPcN>BmH7pu-RtGckXJ_<+^r=)N9%mXHE8H?bFJ#){S^SF1Xyu+XS*$*>D`x#4Rr zp+L&K>&-lHHJd;uo*8dA{m zqlT&y;OH!9A^&BL9V&@nc-|okrqGlD zR@6zUql42cp=inZC5LATB?5A%eKhu}QehoiC&m|AaC_~IKMeO-!_ecDL_-uBlb((e z{c*|FojL5H(6*jFxMkpyfq6LnP41`sJeW1akpc^8^O^Tj{G6XgGCtDz1BNUV(0Cs=FK z30~8VPXyd)9Mf4e6lAB8JWUVqE^UiIq*K81X7}Nu;de)XicqZt5?Me-0EU9>Pw+`T zz40gsjjwq2Iyw8ZwZ8PNWDkU)@XI0S8{key>{(U$S7b%icMEI0I4_<|WjF>g$YSIK<1`SGq8TQuQ!gZuYy%+j1 zq=q(3#}?7LOa;#eY z={y4EK6$-h$c_|CyOHo5myZ7iK+Z(JuMMmS#7j;wl{Fkj#Q^#AvOI9{(0q}&N5E4WEgKf`UqZ+FdaUsfq2yrj^&dl1SxWzrnkx{WPsIgni0eH55>8oyP}G& zmR8&+FOYA0v)It#sfRq|p$mW$-6%^7R(TA^V%TT_QK!T(C@z(EY2`(eOG2)g3b&BD zoJtKc?{PX7U6p_QbpmG^i>p@ac#Ys4fhWSMFUzwWWoSwK5lK@YO|t;H=43K%?ufI+I4f<){d4C5RFZk&Fod*~8@!1d5mSh5PdwDRPeMBSU&^(UkcgUa} zrvTt0d8GQHzGu)(pz+@qw}mNYmx8N`3+t@vi_w0yH|}FVs6H!mqVey$XhgqHFZb@t zFnl5M`qy7x=uEys;_$qH$=NSM1W@KWzM$Frd`jhhvQ(s3cnTC>1Cow$35y>TN)!T( z@yk;e4=*_eFT5~6kr|MX8%CUXl{=pPyrVJeGPV{Hjg98Nt6YRe?VNi*D?@h<$l>n% zHxi_mmhBrKp1N09kxu=vX%2C_vw8Ne<$gF;*V;guYKVQE|6*iY8MJ;i=Ds8M>^M4f zh(m&WL-hvAB+1~|ly6JsuLd&K?J2=)%x0IE9jMV!&=<_m^#xAs_T!+w%YPpX?>=|t z6;h22T``!z489?86sY-@nT8~y8y@mQI6Dz`!62{gSvvxVL&RiCj?un%IO4m1jtA;1 zwK@o993UR22$FWTA_{Kz$p9~=lwW7vJLHS*qh~QMPj3!$Bbo3avH;FObV&;uFjd?) z=!9frCstG}+Iy1Q$l$Z z2X)z(!lr{;7Q_P4;6+F6-vXtDi&u`Mnv#lj?4fk#Ik37K$yX6!(sUi>$scsi(^t}P zQNgMDo<{23zL)_K%0a!(%Qul)v?&06dZCQ6@ibc7fdkVnn2ARMfdtx;6y_f@ItkMT z9!yH$Odx&=26}DT@+)EbAyy9GS9~DX7^Ic1bWZGaZ3cPHb^u(qDVpq<0Cowq%#~0< zDaDWt8b`M%CEyiUrupaOzZF@sC;bvz49~v-2L?ewd$ONS%f!e90Dey5eYw74EPVv023dWIBWak`X5)T;oJi<&+4;Y5n&vkm(!$EX zFugu1>_nB$9CRtWJ2mSFZl$P{-MaN2R=u$le{>G$cZrhMCK#;XZ#4iD<=K=MR7vzZ zA1zhLmoWYSeg@p^H)wlY_`LRmA=t)L2=y0aOo|rhXidu}g3dr#ejq9^tkk{5ws3=R zwCGU$>JHl9Ug{z7)=@L{$40Ck*^ko*oOb!X%w_de-;f-vaVmr~VN)5B8pG0%bgw5r z_eV&D-*N+ikTry7g9x)8S)A|35PUid;a^$=8@Nj{pGU;s8qWzZ)wsUG+j9Uj3K5)O z6=G9c$H!4ElO+At{oA@1I!*CU9dHNdDaL-1qI#w9}C*fXz~MXZcd539QxUQZv4dz$;cX_AA(2DFvjZ7#fnIV zRePr#U1P9m91fqAuMQ%*)Q#H(VY6kLr7@M!;El$xV&K6DWz=_adNRSAdv;m>ylaf2 z=&|WuCkdvD8NEKO9BtR-v>w_-P`;HBYo;A+I6M9P+v~g{Ti|3%uh|e4wDCPSQo)^v z9~#;5mqR%BfR^=k=And#sosJeM&&lsd6olt4aDc!Krei!Sv^uElgQvx&HX4$lhsPRa-y;tvf+WTRKP8Zkk&#ho?%j8uFb016elI)hO)Wa@ zG6f#nV8EwLNyT?RsEqOW+93o{L*4sHF&Zl0?`cXe%2X`+dB01-^X+pTeA5W1JO~hY z+b1J^#t@*e0|x1~Q-uD_c~Vy)DCFu63pvt)wM~XTCPu3QH9h;n3R6cSx2E$#{;^u8 z7-KDLiwtW=3W5)YnkNbJA`LMvd$zd>qh{m+7++q$-E5#R_=yM)b3&C=R-ixWt2!IM z?59HFmy<69f&+aHjdO2I0gFJ+7hkMsFboLSMZFK5R-swZ`{hz-l#}+;Jsa`!5lt5R z4lm^e**)g*3envq+{$gJt~88!&~&|Q{@(2mb`0-$|mMFf4(d#z+x<5I8-y@3X~Qh|79|g|0R(h-tu=8ueNpJQcb` z*EkD|ahl5h@giT}yNLWENFh``YTtI19%&QFilBf@OLMO)gd}E)J)nJIx%5a6cv_%l zGah6&#_Z&>QG1-gREz-|$qk}|xDMWba=`>4qB~#%^P}}b08y3}FHhH$SNVm0Ebq_I zR0Qm;wbCsEWY^@U2NFAS#7I-{J+Tc0c?!TXQHmi4erA*38XGq15h+Z)B$FbNkf0^4% zz>uoll{*|iog7xwU=+ZHI5LL#$bqv?3;JV!odjT_sOQC&xDgCqr5Jy?-kW@XR?!dI zQ&94WU4EBuzs1pno{EDJIj7c20cZWaAwS(v?{~tHNt1n_;k*Z&=-C&^Lhd84Sn@U? z%z}c?Sc)tw)60Q-$S_z~#-_xwzhjmy>;{(a|;U3xN_SQ;t z>TtT<7#AxF+#ZRDM3(mV8S#TrP(Z#^-L(t3ATw^zBI^*wcZvQgv#+&pYe;eNb)KQ zM92j&UwRg^f=2e%!*fV*`LDyg_A&!}a|w<}YPZC9+^J4+*e&?Ncn3 z_axyoZj@RS{Jj4>`co6 z%UO5FT>%{gS~^N0V@#w;^;EMFf?f^}8^38qydnsrTC+vknP~(2he2{?@%zVlF;%__O zod~I;4Bw1T?ee1m#F8O}WLa3^H47a|Mx`$;0@$6cQsXLjh6>p!oI@Wz*E{GHL6Cj8 zv5cy`a9zYLZ3$A?bK}Z`7#guKXX_*60j7pjo>SB0dzS;Cko~xWF00Va#~-hKJ{=zx z$UTfM0+q@J{3Q+fEC9g4m-U0B7f@&1IXm5mIU7A1jp^k2v*aWt2F5|e&^DdSgRW5M zW>b!hWBYKH566)gpXpOzv>z%rSlk8%p+-mY`-Prl%5r0`#D7qK0HFIUe_(vI7arJ_ z@5%v=k7&*;&G}p0?aqqI>`x(-itgT zx%FhNhUy~lvghfs7qu3z8wnl7I8h$U8V5Y7KPSf|c}pYFAOvHf=xjppibETT-w*M7 zF+ibU^#VpnCB^(7OSc2z-ikkaSU zERsxrNl$}4k8OZ`-ogdJbj-Z)=(=^U$H8(hRdGL;7GRcMvF`2l^1gWp9Vzkkc(e2w zsG=Jyz7Y+@AurASLri_Ai}uZk+a$hesvq7*UTV*xlG9!LybqMZyuZaO{KQ9i|Xs08fh7a7KXPI_*fk-!nJwdwJQKgRp-ij9zySu z3B&=PZOc5ktD5VfVYVM%_Dx~IsGC9rw)%Vl2=5>YC$x>+LkbPByoc9eJ-q^@#)0+e zhJ@+xJv+Doi?tz~CQ$d<@UeosRJdv*!+X+8=v_7?>JqsnFY-^8A%fW(~*$W&~N}h9N1JB1e z0J_u=-^UesH75#mm9_T1@;Sm7QYSRuiuaZN{k)@8VcTM`wEo`ah@f2H_%oeP7xVE# zhj;}rK-)nJvGH!S{l%E}1z@%A!dojX5Nj#K+DIRZ;-@tEex$CB)5c!Ky%~h3FR*|m z(2w^6_IM=ZU5#J;akz%C9TDgX<{wLEwx+1IE#X%JLB&gY>69My4pDlf?|k|dWUcSi zS-Yw-OHGA2=NKd6k3+CtP+^k@QYda>Gicm9E%5}Bj95j2!^KYC54Ym@84)iIdO^-P z;N^ap#4!IIb`XxlY!&}VEnFo&K4{F70MYljmqo(-P2JHO!D3^WhO)c#N^~{suGx4B+F|oHb z$iW0Ufz9@#r}=WI;}X`W)GO;ADwbD@;mj=8xs(&OeqnOAb??>6f2YF-wYzj82-}~d zGd>-sGn>a`bXQyJCZyW_8iAn9P_z*_`i9O4{I=he?5|}*Dzx;(Zx|YZz&}K%jzC(J z_Ynpde?4E$!_l_EkD!Ly`Os)hj+zz8^ZQO8anAoldf;CTkpOY1u3m^961Kx#iO*>& z<$C_Drkip575b}+fT3MZ^jw^kBa-2iBK&}z3KlKwfE4cGRaStOsgYN*%W@FK>d!C0 z&d_#-oP*co-X}+L!KH}Md8QZn^A$_Zeqm3r_2Pem349j{@gGv_j9sT??|X6LI`&Q# zXr3NI7m0cguaQs&rF&O(;lV{>%~9N!AY{8t2xdj>;5%7tgw5HA_(Q6A7e4RVwfMyd zlYlNfR`CBl7KY>t_d1IkeDX7vz*POS~$NCzv(yQ4@It<%48%7B2+?NM^h!5NVk{}MkO z+^nXbWpFAuaFZE6g4h|Z58;3rIQkn|kM#$}GV8noLS?=^Bmg)dC|z2S?R&3MK03RP zb#D9b^WT+qHh+Lwr9A05f(tZ732=2(3z@LQ7##Ymw)Buwe&2dlRtMQZTXu5oey`Igo`&Cq$hnLF-a9e?#g;n4TYA* z1@0qIKJ*zU%UAGY*)g{lu%Ak8OACz$8T2T_L5_0f!a`%)d2B1kNBHuQ380S+@JZLB(kckE>o_Z31w=|mDGD0<^vbD?&b1Bs%N`)M z#@36+zA9+)c;Rvb@w?V?CXYFwQAj(3u@CYgwg*Tbow-wV{_ZPJO&UL1EBC@UE`#Gc zi8Sz}yBn5zsfdrZl@{pZ68gRrNE17Wt$KzXPV(HUk>C`I|Ow%r6^hWA z&ht+PW07+vqzh`s`xYEO*NZk{*jr_8e|v|NX`Co5KoX*w$=bbq)c_1iz7Z*|^|%$p zV*)Wl9C|(WA*wL6eeO~Xv0kt{ca+wQcs2Vvz~G=KH)ImN68x!m(k_&}VWu*wR`(O! zy^Zfjoaghp+??Oq{~ckjM^*gXt_YXzzQdA%=R(S#v7BiGj85iFyVulVoEvl@0g@+c z8T}#H)-~L_QlN{+1=-yk{oE8zRk|n#2lzaLuvy^bfn)K5Pv5FvDd%samO4C8Rl94p28>Z%cHb3~=9mj|@M6@)6hEVYc zq`I$=&qsN#oha%2VyHN!RfHnqkc=cp?1Q!TXf^Wkr9KS*hA{nSgitK6&*4oM45sX>T0J0CKwSX(^6bnAYA3eP zMCFs3-Cw z%)vE|xnljnH#qf#QV_ksM|vceec>V6?mo}f;W@l{`6bn)>MKWv`Go5Q%;Sql7dO=k z>IO3H-%vwt>IP(uefGK_feJoiyc)W_L>f#k6mf@r~u0AE>ElFCE6zH1x9{JPfFr!sd&*>hu<*5+DJ< zx`y?)?@CvsN-pyU4lA%gztbAkd{GqA0VyEFQJbyKIi6Pse#1AOr5VV(#a`e8Nua!N zILqf>{?#rLL^>ksNP-a?-&HxUB&UXhtmCQzD&_bNPZujBnIX6!g)xAPMR7|`FZa#q z-FW94{b{bZ4WS6NY&1MHQ;2wiZd;58UgZI$K$>$`dUK461Q*X431hj&0$(-l4TtOq zJZC>3oMjs+8!$`Q=|t7hx;>$TJE{s|-YEM^t)i#}!~5b{Uti!M*x#Zc#@Xc2<54@1 zR>Yy9h*!TI9%u3;4tFF%X+9WSBorgd&%=YA%|c&$#1Z^1YH57DUt(mM>!tqqPG3_X zA^49)9BTQ9Xe;ZHc%!n&lCyNKzUtrBNrh@3{)ik%wyJPHpQ7EBh7eCm8BV5QJ?OV) zbduVC8d9>|&)A(s`Yn6S*^+HQQlI-PWz21Z7!2TdNCZ^H?Jlo_sn^S0M_djN;VBoq zRG0`?X7T}Vt_-7nTNtuAeLP>!gnfxF-+ZJ`M0a;$55qVzw z8h9yynPHDKe)qF?V0hqL!4#@0q<&WlUqOHpv8S0nCg9B#x4*&%a`PIDXtc+ zc!T7o2nBNqW`JCwitJ&1euWoqTP^jSin#h0t;`RXC7JBPV$4($kp&6uQA$$Mmz;m| zaj`&jY_yI2MN#lMUdM4wNp_R@G)v#waM{d_ydC7~LBIAA&908hYOa-PseKXOoU&2H zn}W&-UYVq%n)6)TblPRT6!Z3ay{F7ZM@0}(8QxqZ&l)CT#GPh~;s1m+;)Ni3<^Viu zGI+Io>aN%nj3qWx&eE9UMz_}8h zd%?I}F5OJmwG2QjU~1~ldwf=W5fakjAWvkKNnSX=>1Tey5faqGcs>uOgbbiKTU)8J zvg2ZZ!S>#Ia~wAksA2sg>Gd30&<^7bz-d0kt}jh=?E;*XIvw)FqH=(k|1x7hj}&yO zu-^=JqLWNs2%HDtr5f?G8&3p`=1qKe{DlQ4bHk@w3a=-G7W*+y_-M}jT%{(p&5LkE zON|-FP7uviT@VHK^>q3Ls>Sa{-~@0YhATp4^Ai6s({qDB?gPML*8BTC|Io7iAm~xW zgpsI-=TR@QSY4k^kj8oe#L*-#^ts0wHaQ*kKFwF^%ymL$gYJrzxY4g+`yvYgV*giM zoxx;U_m|H*#)U6APMjQ%`$<}AoQkAbp9pwd%a4*U82w2ib2e>dQzEQ8L#crv#lZ%Q zFL*O>faH&6%4_+m(VRlwy!}o$Or4bi($X7e0NStuVj>B^V165$$TQT~VrdaTy8sz# zVJ!mxQipd^-e@0xKzgd{5EkmW59nH5qnODjox+X^JCe2kz!mFNMVak0=C>JSCHUo< zaRN*Ww76^ejwme_FQygTMdBNo=HB2~3c*P2)aSyt`nX+26XNPn^fAJkwEh+!r<34r z#04nkJ)2LW8+sLzi+NvA_aHOrgrUntCqa05-a%+xZ4D5WyLEo-^rsK0o=ka$SJ~^%gFmh4>?<8?y^-i?#;sXl1h>Ti-nBPDF5`a z*9=EzU}`%|r&nzJ5D3FDLBj2NO=gJaFN0skzSYS}RPJ4NzrHX%K|(In|9N(+6aRTq zNI)jet7cZv?mLKgp#YL1(tSUY^nL%v;}QxjUJyX{{qo3*7WRBpzuOBROpD{h@P!%c zq?V0Z-CoA-k_U9%^6P;)iiAW2 z`%wj0yc9GMo5HR8jQg&wlSl0h8mZWU8E_+x_2-P&v`2n&7~qJI{(yMgZKmf@lw|*u zsO^CrtJY_`uE)3hC@ddrG6xupdyR*iU9!Uo%Yv+&haAek6zLCOp5{r2_w#_Vem@DI zC4DB5jGJ4a|Mi-fWz~;g^;<-8hsSeF_0fF2tv|BT<>8AMJ*nzrO%%L z(JxA&S!GEk5Z^q8`|Er<#3wpq(c#)#PPrrO8!>);kca$UgUW526@Pq$EtDf5fj(~i znJ2a%5Iy@YYj;Q|Y&Ys+h!11%Rjm*>^oZ6+(T*=S`c2$04A2|}++^)nK}WY~tL@^s z<#Y@20Lx=a-d|b5(cW8V8h`WBjllh7*XM*#zQasY_-C*brke%!iX^C+1;aE%mK}(& zcf3#B)p|CZo!LH&*`I?Q?>xvORgx_iIU&c_Zq3QlN9ZD2VDJ zS`R}H3zB_Ok50H4;N3G03O*spCXpfzUyQ$lI#2@TP@Ae8=UD!IvE*wuN53eXj`CV% z21L~^Jd?t6*-z8YekRvf*`|petEdQ4oH=X;P>=4CByDz*()VHVJme}ImOIhPL++M+RKks^vkLYo6im>SP$X63`1x>jCUFI#G zWDlS(&LiaUoy;+XIXsNh@D@lpD}R>WUR$>v^InL{3scO-e&<%6<<2}5J6cl&41V5Da0DgU6k_y2!oxsmTiy% z2*4uqN1@KH>g8rFh>|&=t#rn`QS-Itl7=0P%|vn!I0fJ7!|5qiE^c$%F6nzog=W4Fq!dilE_e_0+n)9-bli8k@J`K};tL>v^2qu>en z88;IS=wGf&;%M4!pjwUpK8ewv?`p12sIc*R+O$K+g6qY5d!X+Qh3P#QoB}vtQHnLS zmu9k#Nd;+;RP4po@R|jGT%Rd36Xd~Z>QI}Qc$;eqp=S6x2A^sH81}@w1G^>%lZGSc zh|DceFCthK!(m_U{(ecI5wtNx@Cci|dsqMi7-7jR?VD|_(V>XJ&iimbq(HcS=KmVk zJ$)ayBt@b_`rKM~aK1qmMx^E+U$~_-NP;Wi^rZPSw>jX4)$3_pT*TZ@l(`@S(k@rK ze>Ur}RqHWbq3J)>2;2I>^qc%GD~O;hx2p4O_^x!2Gu~tcH+t&L_BiS^%dx%1=ahic0m?PU<>Z&+Y=Oj$N7tcNm;|piofYb z26TEj8jeJjp9iGBy^;!0QhHiH%A&+Ae{KS=QPD}f%?Z_gs(gyYLmXnLYZP?FiBcjy zYn`QSRpXyNX^RKVI{V6W23GKBL^ysKvj3}ilR-#K;7O-Y2!t3<|@BxH8WC3 zaqM4}E4Zzr+jfBvRNu=G-o3+gVjmY+cNXFF%y!+2qeDNVA8kZqF!k+4fGashD7w)7X55gH-b(?5WLt*lr^}94 zHQz5c0b%V{H+CjSTF)9iPV{|`m0jAOYkHhFSImr!c`QwSXVM|YexP80NUkn?9nop) zy~GHH-|O;zpCi@;jrE<47>)2vufbC&bHa&zJ4>S*8X^1CDP7WDe`&PhUuF#r=swH# z+vsat&^I0^T9U_SM@L(nX$~t`n)klk&eHhS;TFPe0a@{#jD|;!bxAFu57agu9aq5*I)Vd<| zUPK1gR#s>%{^W5lSFy{Xq=+B(UC&3kkP2`5IT~65joz}&swlIS+(_w!gvlgj+=>vD5|oeMHUAL+b~=j zFS3C$J*dvf=)w|Q4Tp`9(J0QqgQbVp$hHxIDYU9gU}wNgL5=%aNB@g1O^`tYxQKw|jWU^BA*T_5oct9bsz7JPZcjH016PEVeZzA^AGmj9-aeV$Iil24` zs}T3}{uL!}_Yb%G;?d^6-FxTObCSx(GBYvlYlvmJZ-MJEgr_y z?2p>=^ZW$EX?H_RbZmo@=sXM-2}O{naOrUj6KBF>(g|osM`+KWCd5-3+cV(;;G3)H z20xZp9Wk%v5g||g^dR&|4c{5{Z=qCUL;=9Ltx8Gt;Cl%c2~hAWXcy7Axih{*wD+hX zz%e+zjX*(gXvby)jOL&3VS<^D?1wpT-;f-LR=0H^@W` z$B2b1d*_V6toN+DxJgr!1ZDywA-IhvjX6>Kae}+*(lzIgo9zOtx6yTpdfFOUpCVXF zArSZ;7jVSz$D43Bq88Wcvo7I&XKXrqfV7WH;B#<{yygH6(@uTbZ+#}pEM2U}fIFSY zq>e#f)_(A%=IZZ>DI?zH(7E(Q(gSUM1T9D>U&EbAR-pMHx!e&BR;x0y;4dJi9yh%@ntiI7~Y}Pq+Ou=!kh=eWg%e# zd_aHd+4N&Z?1 zv|Uc(&rUU{XioPU4+EhMHKo*XEFQDrL6$sh{W`hC$-Jr2BbSX_{N8z=)5U{+Yc#CG z#;1agKzE3lPZ4ciIa>peust~;(;e8^+M6si<~5uPM0jW8d?~RU8u##c7k=-G1Kol5 zhmFI(7vgIGnf^ikJ}x=~+;auQia~2S{zAkdW-wuwHp^b)v*3YGSr_|#-3(~0`L<%U zyszoVf5zV_r$IQz#lHDHBMsB_{3!u1>YXmh)gIqb`H*tSZt}vNKn`WxGfSca>PDaX z^;~k0WMZFBhjIeWonM*kNZgg+yvD=Z`+Ar6mpa^-;0S=`8+Af+a$L*Ye=2~?&NRkT z@v^X}7xyAoh~CiTt;FL_AVf=23XguI-E>%vY#N-BD#zF-?z>kkqF`uGXVc2I{qBax z@K@D9?J^7Xy=W1G?IUgB!a&z3>50cFzSw}~{chTFUZ!`M2`>LtZH*`ggytZ9((~Wc z(L5HACLq?_&i0)E_Q-~i9x9M&N|Zq~9X}RL$(G4W&sZl4hBYF5N;v)mW97eHdQoOs?@&FD?aEnPt zLH<+;sgUEZKLw9SU)5Kx)Y@q|Yh?R8%-&uQFEKWM37)~~nYaEL7vbT2K0t|wO97=6 z$(GO49n6upxlbbgh~(^H--~4xMb5CY7VoS`iGzR-e|eD6CChG4)sQcpGRRkv8-n=s zx>7SuN4^531MA!Z^wT8M`Qi2mzLz_8zgUMX_JHPBf+9%Lha(_^t%VS`&imm0=MdVk zUz+A&zkZ)iTXw!w4xr465YyAbb`c;8L>~z5ww~{RHV|-z-yce4kVlhBDGsUQ!?;%w z6w&XSMvtob*30S)TUQ0=%1!E(v#tri`T6rNt`7L82MWyy#I8N#n{qa?hxHcIw~PeX z52Pv(CK5t*l&WCIphK%y#I&XIH6?92m>ZgZ+Re>*@xxSa`BCj?a_=C{ zb4xBLtFSLSEZG^(#cj{zGHA$GV4d^Je+c9q*!~=;4^L+atQ=wjOGscs`dzNS`1Nd# ztoE+3*x7N5STe66j#JQwZU5{F;B${31fX)dWldX^Ipwt9Z@D~>&X0=kJU#$LkOHpq zYcD3P1Li1%PIL&$wIQ2zko0=NJP7GLpi(?fSiguh;l+vIo_eNKWqKK0mc!ck&KDb(eqx2p+9i#3=AV(;y83aKN?48u3CHPYQJGo#9vFwe`ias;@8Y@%CZghF1)2p8KI?-c|R)y%#97>HaKXG_xrDYxj5w@v(+Ywv?b6msBS zJc2~fOycUpzDB}r;_Sgz610bk8ix{xUUlZm~r7V7iWdEVdmeJ`BsE=&?Xk$zKS zA@y{S@K@q$9xmD=yhLn|d}+amx6Vj9-nBk=({BruJ=gqda+##hKHJFGV7#CPGmerQ zE->VQd#F6KGH-frmP=!D1J`Iu;!r0_++?jCBF!9 zPpv9cDz@)(@C#HoB0F98i-t!}3_sludiUzMV~?Z@TpQsH$W-Yw>+4$I_Xidze5BkJ zrIWsbLXnKh1HsK530!=A`)GlnyP4wN@6i<*m%k7%1U$C@T%mY~=Ff_0J9N?8XBYLs zC!Z|~Z*}|Qeker>_h+qoux0Fu_F!(*Z>pAxlc?g3eeF|-+`Mu{*3m{_;7yq8WXAb< zIR(HDrR3V5d}VW@=WV&>Wtb&UEY)S2y~Qp-oAg%D6r-C3>2OB=%+x3Jpl|TL5mOR& zP}vU2ZNe$Ua2N6AF>7$Y$1~MZ%s&#lSM6~ z(qfD=S}L?WkWa7>bFTt4Rq~0^D<6EbV;)oA`ys9MlqRgYg zm(;24vvO+y-ehPkLi@XoQw%tH;E6Z zKq67<+*(TNj;eu34_|Jdrwz=@z!6&Bg@r@Tblaqw9ki*olOQM${&#TLPd`2o429nl zkrD)Wkp$824TgnHv*&&x@nY_MFc8W zN(!NnQE>)k+v>ivdb@}>e3*2bGn-=dOWPy3k%MaJ5y zdIsf?EUhU!q<0>|7U^UH?;&;EV9-cYD@*y|q{k{R>a4k%#ZI^J!!{z#l=v?|L~vM2=@A6R`R=&$j91DVGlO&of_{6c6RhVQeBz;$_^VQ@IY zXxf>d1CnmQzhs=*kQ_rTd%JBgq>XG~6u>_p|18KXc;q%sB5aD+GVViewUftNegY&GUqr z9sDq!5~75Cr1I~ar~9^g+AB8%9A}^~8pkSxD%2X?mcW-xQ(g+rnx>z`=#z<>jEqQ;1q;CakGa&x)a%vhRz_E_fU6(`J)8A@R~c3UR1^pltqpW z5zy@Z&MwULf*F#TL&^P+6qTP3f_E^9%K<$O+i^w#RmT=cC=l(9h2i(+Cx9BpYLWe zO$M6Q&p<>ehEco*DIwzG&JJ!FRG*xd0(|HTxU7lpnLnQIsv$xi|a%{YAQz(T$G4EHW4Zw= z#>zUv3z{7KILFhnQgW;hgH14^+Q#sn9kkM@DeaEOT`~X0*M&S6;*qP5)L4q_F+h|z zPH0m9cn61fh9q`PTg3g$HcG77pI+g_9)$AfwESS8CVT9`?4XE0x02hqb9@f_L_{dl z!^O%e-N6CCgU!opegCik`J7!qdw5yM??7f&@LXB1&03(*o>)b>V|_J*^?hOq;;vo? zGHiN#L-Q7!!fL6FJ5b*fjn=QN19yw2u;2lTq^Hqa6DhhVzXl)b1_o4=n+QkPNP)-d zWyeT?&rwy_?N4&nMP78M(Wd9bmoZ@xIdRb>xf#wpPlJPT;2J9e5q-1Q^L!#Ya2B%* z`3R~2{*oE21nT+w)W`mW@CP~qkkU&LxQ%>k@l}sHqTuk;vXb$ISV1?Gvic!|dA|*i z>CkhO#J%_m&Jg@Xch6oRMB>|B1$@{+edk``p)GL{kQD4;aG>%ccK$47UvmSl;fvfI z3fyfpVu7}Af-YDPdfz*qx>I#>TJ${OO4>;XX5|C)~sP9}%rSMI_(Xmd)X*TPi!fLSUd*gf2 zWr&fV^=&C~amEoH6jBIxoCMgKN@eUmnw9W&$^l?veO0+`sbs2kM9$sYi1h51#J=p& z#$`Xn;1_BD4>cBgP`*E;Dw0ze&%FHSz+TQsBCQMG~>T2x8!FDenFXR3mf}6BLG2ul#T#|Ha>}cH`9l}Xsn0u<5EO2JO#b*+o}e8|i4O{kvpGY6nJ zX>c+IFbD8LqW0cg1SUKyh( z!yr8#2$@A6hDb6GG~HEb%?*o}bm*IxG$NGx5+k*?J{1S3UfNJ0`8p^&2tR=Xz4-@p z(rJ`6lycv$5l({XUE`_Vfr04Ppz?cHbjrasjaD?v*YhR$N*iZF*YURe5E-g5=!);b ziIG%?RGDfkB5X1kOAS+|?8_UB>K8~WmFO|ZKcX?3gIR#aOm6D_0!~LUH zL5J*a1<fLoE9^n(=r|En zMTN24#|Uz&&vPu^>r-4dT-KTciy%0uVSI({Y@gUh!`-ZkKC=`seR&od`5-R2ne|>9 zL~}3yIE4i>nf+n!a53za{(u;W-x8O=ajqQnC@C+ZJ#=2tA}VUvBio;z9dtRk-K{-P zsg&0-z*!3a=5c85b&Fbfgft4M#4Xd=EBM2Fy^@Uq+s0Qa)?cM<7ocUaaATjENI;0Q zVs$sTaf)GjU2~_{3yJlVXiCcXG%zE1$+oXMh+P(DkB8$W{$K-Q|ClKd>}9P|wdEf< zssl^lDxAMRU_F+*(a}Mw{hmnu;ybE&C*t1IQ%_%>*-tD=`^~?=#;Pz@JaohsSj2Y{ zedpsRd`(Un^l5va1_{o;^c$VcK!u+XbV}&IaYp{hRhE>uF?xAOT!(r5N*6mL80E_b zr1@>eGu+49@SOII0b_H$M!FJ?w4QpNthWC>&acl6urKGfT6T^IG%r0uf;rP+f0=4! zzDCqR3qE+;@^9M$dH8Y1fmjX9nS=K7n|*+7eoUZ#*$o)b3w?XNsml=6O~So0inDJv zccs(z1*H}bH?(WxXMuIDkZTbnfCp#mnT(zVmklRJ^ZRCP-|M8Q->MKU+CT#Tcho;X zYt8&!!HX?boEO;4`GNG2F3g}N*roM_t5 zDnY*^xzCC@w>4P3biOOmej}Hh+EbR`FX2IfzBHy5BWHrtB3q=qiBJ+R505_`pHNg?p#MR0rRW-yTk?IE-;tn!ZjLTAFC||L|uXe1ink z75Tt>w$WdW1}TB8FHeXf{JL-dyFkfatF=UZ`5(-c0ZE+d`S6krq##0Yr22X4-!gd>ek*cNO3v}QWLoz z&Ez$THPo5*eO?lP6-n3Ap-W{h`awOkJsqDP!s2H*H*`;*$S_lLnRfsOm)LBWd6&r; z`;V(Pt>k{|f1pq~j$?3&4@-RL#RqW`?~e^=?I~=|d;g2f(+8?w^Bp9T zNv{c#0ueDxQHbvY-Rks77`i<>JZ=-|=X~IS`-(b>6K>C4)jnVA`_9j*BfLg{2TmwtWHJ#25dHPZiSj}TaMrr z@RYGmbaBvomhEXiXkwTGkH8-TQ{-F*hK4IOe2Op9$rjWe3tS4?K%?Mexhb4dos2^z z-FLi>E)f~w)LK~~4XIIdDA19JLym8RN+8kzOyDqx)aS=Q>Lr34bsQg`bm?EUN0EW-27m7of3_%9i~RI;2qa0*H57>Z z26+uG22WGsPhu^pHP7B~C`PvmzG0&mIWn_9(JlP^te(b7)11p`uQxAy1$etU0H3|f zCehl?h|Zq`Io&m8C10>HJ!@xwfDJ{Cw3X(zm zzYSNn!TUi<;fAw2UMc0V)BVKZw#s2mw zoTI(2xUcYFpKR+@m=2H*&0M03OtxuUAhtqt54nXD!ULlx#4XFuSExN!!-wg^SVY`XIs}zQMESTwy^K}lWO1K;zP7}+f@R_{<7KI$Qh4f1P4eT zWMqHsZ{L1NJU^5`Gc1jyvsSL7^Yki`_4Yc``v)hP1hVHhHa}`ScPX$LkC@@UN^GeX&}D_X;>F+3!t;(qxpe0^cv; z{`5!}Ylg1|y&Z61>dPP;k+ULp#!fF?6?)C-0lj)7K>RyM$_SHh_qPbY0!sdxr)qCH z__h4<%(g+0V21n?5d=hi&08DFnXO;iB;1Z9lwkPrBp7z2{aV4l>QkKH6gX&QgY|)2 zfP{5z+v##fe^5XPPsVo9;yAz^&SIac(3ArailC~XKyP(OOf1TG(yK7yULeU84GcwK zX%208)#7h)*<&VuK#i2`l}0ShCP+C}20{>SWj+jNk~x>>deZkp4ow%CbvE-{DAC(` zaPHQItqT#-X%*uIUJE;@yQM_74`I3#u4`_{aC$%GqFgH**+4%S_lqDTy`--B0=8wmxK;YR|v%(n{ zZ{Ybr#UnhX2_GV93SarXTZ0YJb>`s+!f@ji9O|5}ddOBQ6d+=tTnh4tDC$6y*oBl_ zN|N`(%b70Xx>G92v^~fwc;A8&p3N*Vqc-(XU$g+BYsC7h6o!}N?H8+t6+{bkNS_w= z>IX*r%GuUi1T)}OB(A6t$i)bn{m^aD)W8d`l6-R{a$BSf=M!mPl#gBUtt&U!E0W}) z*he1CzwY@7LLrQglH~3*?!P6UTKA1@5i(R2XJ$TN^Yz*sy}IN0+72>HXwUkoDJ9ye z^wUM&(dmX;fRib|crg2gtgHKh7}G>=`2yM&Kpfinaj5&vVtzgLfd-=Hyk~0*!NqX$ zXM{}wN~nXIcaV|vA!iMh`F^UNeR&nn^Rz5oX^R-$#+>tE5buJNntt@KrjhGSu|}n^ z6x{Gtdi~2M^=F|p_Ps;OW9^;sTxkmyPgqEQOSzLsOuHSN6Svs7muSCXT<31jXn;zj z$(>DhR`YYO#zj6Q$!z_G3JN#hn6U0;yZgQXRMlKkd21$dwvSl!kk*w50}X%=0l2w%sjfg@ z1qhBj#L-8b`3d=CwepEL-9{DOZB;pTxUu%+L~b&j&!6O6VvsYZc7kAt5^>zlerM0G zhzeQTAo6hM=w@KFJB~O8#Rn6K)@je8x@}9J?fqH&T+Pv&D|%mReXY7w2WQ*QZ_ywC z&52OzrWGV-+UttH%t~R3&}+^SXVrq5Iy{?>)gfN`3l`abc-Fh-N<~971?uZ*x0AXP zc@Oh|Sh-b0=M7Dcio~GolzwMp?}DC~+y;idc^AH9JF=|;U5NRxU+Y%~^k4Pw6uY$-&$>!dqkkKyL34Z@2R z7a%owH_l>KnmHsNCIGV;OBna!NK8`_NkL$oDF^M86uBdJSSKk7Z6umDH&U2R01uhHOxN*5$!iw5 zt#A8|d?&QT6*hfZsf%kQ6_i5K{#eQ5TDUQ3>c&le+CZq^61k?B%k9PhKZH&Tcg3Nr z`KfH^V=p7|1n3oWFPg%_&FuDaM#~g`4RfbGyo9S!c3FcR=@Q)Vl8AKagkmF(2(YH#rC>edc!7L|wDs-=i7|h0J-R zkIZZJ_S3(c9zsw&-&*{#L)IFRPp5FCBv4M6XDRTGbh zVv9yV)LuT==tcr7w5G+=)7d-J&2@e`^y4bui{-8^m@t+cQ5e>_W9QcjI49}UNr8Y} zfdAp?eGSEa@rEW27JiwC+2LSRSOJ`d6hpeIAJroWhtq|m`sw`PP9GU%XFlwM{t7-E z$0GH-2JusxuqJZC5o`?X5yl}G+X#Uc<-JM8gZI|O=o@E@|Irdm>Za>vo-m|zc^z#w zO~2tkNMm}B`zFPzj|t&Otsf1t2N>HR+aFDP=h0|EQvY{s!vVYZ8%q3;0HVF&h`WwI z>LParm_f+9u)RBYR#I9mgIrz73%Q1~|D7BXqF_(yn;)q;W2Rq1q;CsU%G$L)o&^Ta z?N8zX8F=y}3*tGzN`TO3nCa;kCV*%Q`D}_XbsOcmP^odme1&tk_LGzVDkk~{)byL5 zopSNu`|qE30A&Rx@pt(|M9l-#so|Y7u3lmHN2dqLsOXJKHT1cdl|hB}uJ-L#m!TS4 z<;UHDVIRhIlE+nxD6Z_li~ETW(0!NCNRfm|ZvV>Fm6({LCE>@$YKpL~-_Z{a^Q(`| z_hHQ{#$LzMUi6>k+)SVQ#?{&1&(#fIkj8EmOdV_T4c!&7AoBAg*V+JDS9U{tv+(xi zfhG^xX@U10gKdgHWO zS8LpL6mOT!VVW(gSVtCLbh<%yhtCe?T(q~W>7#{18B|69Rm?H2I98r26M={Vr?a{v zUQX+Yji)4BkZvW?Jqv|THguUgWz6`7C@^F~Qy1~foUsivtRg}UtoG-`$`Hgtt$gH3 zuJY-?KS(kedK}e7o>PAOS4)^D z@AZ16dru0py+am0UY z@QN!Q&kI#BMs>F|+7!Vd!{sYt3uOoO6TV%1=C`Z*#)Y{m_dZGqXi~8$NcWrQMsL7; zOg&!#grC}{Pg!-Us`r#o0W<$80>o+y-Za(ep0!Igc^a_y-(bBG2nt&l_8VnW56=o- zITV15(mTjMKxz%E98>aimgMEZ$1U|vC%!d-c&gkFAqO8vCSZ#@PX93LwB6D*UpD1vMceqoSo>4CAyJ;>xvV%aDV!!|6EF{nXMd zaIa>39?dq(M2asTio)YjF%Q#v=GF__U+6Q-oMp7Esx|!@6MV|}*(Y-F?EM1)(+K{M zop1ga>1aQk=#RAcZ8RvweCNpgS4uiCU}EWdb%)M7kS^HKERi~M>6**D9L8-Pkvh>!4X8W?&SUxl-5zX>2| zTXga%-4svpr4ATaQG)$={|W)DwTrD7s^GXDlVluZdSEzC6Nm9hc-ICmowL=1gtVP* z)g5p9`c~P16Gcl^gvFXIVj>0F!Wg{zQcL4k&-r>4A=2iEF~oWJUAzaksg?3(KP`xk z_X{Z$6=8m3;4CtRHMYjQd{ivj?RnY$`p+W+m5GUGMmA&FMnPty7~nEivbpmQ5_<~5 z?d>Let9M+WQ|Co$d1lV3Y(?k?$#lfOWFV^98y{tiX%u1{w@O_<*eq|DW&pvgI&h?F zOYU&fCBd7`O|=&9*zT;e?lZ*D3O zAiOX_0`&CN7<-*OSrJuvD$1;wd+oV2hCKZAzxUR@)}+55eM~+_%z2RN)Pnz_duk1# zVwGh#?ZPfY^^*1;;%H*ISmdqQNHc|uS8u-&`9rcwmwfB*C9=wRgpqc#vmm`3%a2a3_%ot78J5kS* z5+g?a>45ew>|Ol%f%h~2SQ9k0v&*)mr9(%=I*&OhxQ}?C0GPYC1X@O#jsWvA0S2Ej zPQ-p?Z_ykOT$i43^$er^Fq7;2m~>&+YEqJY1e-+y&mE!&>ISlONp=Sn8em|cW{(MC zxO0}&AWk%#qx=Rm)n4`B&w>D|RPYf6=;*I-$-gr6Cuv~4qIB+p$O}X`C`^d^f3#3P z(I0|ANP?p0I$`009FWi1CXZvNRpi zX1acof8YqtcL7BDaFp`tTlY`Fn+o_Ydt|Yu>jch6op(QSD(*p`oua($&(ZnwQ+wZ9 z9}a|--i>B>-7^-jiL;PasQB7FFkGZd&V*Kb7A(*2wp2k*@L?-?E z_y-!m?7g$g*YC1(`DU+)<-DRKr2v*E-pGKRPuDDOwOV+*{n-C@+W_s}T}wKobcSW> z29WZj6dr>Q0?H2;)P~UtTuq~;vGcu6_eZmTX$Te=c^A1}>s5_4_3EIW6&c<^5%Dmf z1safb^YHPY-72}rd6peO?}tKt1I$-Jxp-4g!g6fgkn4#VI5#TrX@B&62SaNdP>>(~ zLUJY_Vv)M#O4`+A>pFjJAL+c8t^rTt-@-z@yQ9A4xOiH)6T&7RyL4iajzXc zDHK?(;hEGT#o4b!HW|yB+9x9UbVGZ5Ytldg%%cNg@*>w_snP)tCgdIz^Ke5@v)1^> zeMfwyho4wm#BiZJ+dB@A<4F@K{D6D24suC z?X}9n%@*5#$Zz#EmxGD{?U@HC#W-ltX0o3+Iw}M4h@$HmzkC@Vl`5ltptK?nPy1?> z@WY)f$if<|i|rP$_-3Vb8t~)>mB98(L=-fL4R2yEmpWmB1vKs-rI`V&Ur&FZhx`ZJ z*-K;17JlDzbr+GH<5U8#QD6G3FVFm_#CZCud95Ra+%FPJ(Z+nu5vngsSc7B1rHF1< zt}0XuVHwh{p1s(vC+sT`sO|AC2Q=&shws~TwY3DDgGJPBzki54)$(7&k`&Vq$aC7U zZAg&uQaW-9nzvMKKGqSs&_R#~_d4zH#Zl4ZdbLR2*4dNFrWb7ggYnk_%EbrD8G6IO z>ss;`GXfm)d?S!7b?mh#1n6ut*D@6f=XmT_erz0YJd4{G1X1FfgxR}0h4DzEEy zBZ*Nmeq;xO%rkAGTtW{+dG4F1-`!*3&j-CAmXYU0pMDBj46pj~V6oyb_xN2>UNb`y z;{&I4lFQebeycU6#VzC+di|jxm|x=G+u3{3uj{WGV(`7l7wfkeQGnEVCyx~Q9lGWa znc56BYH<2OorNgZph=+Ja*j)VtO)QIHa|mpJ3=atHZLO_-L}MRA(f#it`D3#7*s8)G5b-ay|2%2rgpao}rRNy! zK^C?a)?x$Cd3Ym70uq3GVl7poQ}1`zCqhVY1767|-W$&-(@?#NceW))7&FAQ_4sHBxvzx`Z>-&|U+>hGdd*qk?D&*0S zGV136?8(<(v2%HRY&^3;bvoOC3|A;z9I4@D3r<+SKlfklHyyRQXx0ci^;jlO8Lhpk z43`BiSFi?3XlKMcq-PvFclZGb>Unbi*7}m0NZ~+Yy9${V(r>CB)dTJP;LAKfAdvVA zwp@(zt==nz=uq1;EZ$RrO=#!9aFgMD6!q2kUHOIA5LgV)hX{}-d!xZK+?PG&;d8=^ zyosmRNBP~k*qsMQJ2vPsFd@L&^m#8aF!zU;1veS7X*s+FYCGlwtvOa)gst5PDImI3 zggkXB$Zc)5Gs2W?PQ@750`=wirK)yW0vsfaU}Aja!9fiW03LV&+2HfIf56TNh2{Ra;WyKHONJxNLzrJc zU#BNl8Au3z6wE{XH5EdNyXHD7^yzu6wump)f!C2*F1%l#nGEvVuCF$JV3h)$Y+A5E zC|C&1kOjqRe+?H8{SmX+oCCICpPp8&A56u=?L{nhaaAh~OQ+Xo+oQRsz-rzQQIv&p{u7XGNPKN}|+Jb09PkEajQO@oIh!SKhm-IP#Ss z$Sd~o2n3@fV6dNb20kqZiv|zmIepJ2?Ljneg&`tcalfWr2=Fo?18SpZ@$WlHwCdT+h(DjVMTfh=^zC zq6lHihkp~p@@na=KGmzBVQhq(vCyce=hI4ZQqlP% zCpArGkrRoDF=8m&0+Y>UG4&L!RmH4N)fub0Zyh%U%?J!tkZ{XFQHUi;h`b|Y z9zt-}R|FmMl`U{)|8Z0^gyrPQF{Xv`NeF+AgEU?Z;1&*X0rcJeWS>sDRMnUqAF53< zOYW08q_->I!pZdz<#UgeTQ-a&Eg9AtPphU3FJ=WYyg*C0;p^AZjQFg~H6%pCFjCkI z(D-3EZ?k>JTP}q8-G_AeLpKLQ0z^^D zr>XjDo$|vqh)zE=d&f8D&S3x)n0W7z+I9J zni>gdtYN8FE&o)6qKp)Da!CiLWHzOvujco5{YDh-`t0ZH-5C`ck<)&VWu-S$pA%3f z5udot05nhx1HUJzU*dGadNKrYIgl2^*=>x=t96eK5A4Z~i&|AyWyAao)CmOADr6Vm zUcG6jF=KPkqTM06sF);}cHAa!G^&#FHQpajm+t?jYZR>|gaC^7<`;u^{glu%bYdK| z#_ju*pLlZwaZUH^b*aYM5586`dNdtsAR7E`-y0SL0eSWS1^u{&JEow9?L4?>7$t2# zAad?4rmj$y=xsXvdn@6F%)9m;diJM?4J#zN+-{#=^B0N=JBq?^>0`KwPQ$kf(gOCN z83J>4Xp@6(weiMGP=a9IDC`!u?>BN1BDY=?a>(F46;%RAEq-cAXQNAXFD<_+rCqywb41D^ zK?EF&?YUB7iNx3rq8IN2Y!YQ$8WfxG)VN*y4lXx(R2BrxxtNW}<^|M30bA_b7olOf zznnC6Ilw2>!>4y0GSdgF{l;0?o-rne_15d;QPwxVp~{m)yICs)m#PMDaK+h+eoH8# z)BF=F#4O&{z#6y{7D@E2-|KY_MSgsIs(Sz}MWTo;A|1Sv98y*eEjk1tJCuU%$X#n= z^{uFJbgwpF=11YV2Ri{v$i8%cFXPqiOzOujC#fUW2r`WO5;nShhu0rJZqIGQD?U|C zYgj!D#L`g}ribEh2&Ox1K+AI)#=q*W=x8+iqayOB6ca(Hz?tn}*iyMZpmV^Kj_iAY z+opmOG~Q3&0APV`~APoe_bM&|1DD)0~3D$3foxBjW*&Pu)AL@-#+WOm0FiG`) z=0U<#`y~wK-I`wR`wE$64FhT1EL5vFuo~=V<|0{FbXy>o#0cqLoipntzsqcP!cS%n zR~O2ryg}4Ti2Sm)KLIUcdbcC#4_=skdBc5C-1T$QP!wk0i!=3`qO7YW9rc!2d5O3 zUjK;&9}ARbvfBO)EzFgJqZhG;CM=lw*O5f-2t*PSc@0>?-vJf(6V8gfU1xZ%NMG`4 zi||ZvJJa-tk@I2gJ`5^q3>OA9LOqYEcNBxlBY?l1oNmMuVNKr5xhDiNBXQbGAH;k? zmL^y$t21YC$p*gn`+-$#T+>%yc_Q7PP3%O#E9H9Ws)Indf4~BA5n}H-_Qz0Py(XiQ z8LXc2TNG`oiUSxDx+M4X?P-6-51yN*{A;uRXN% zE2nfLmp6-QXW2uA zfU|X}W+dEY& zID7>$%M_-N?&hDHE&&{=kE#5fWqNxpwho?6i~S&VgJ8_@cH!yeoA>ahzaHY_^(>xb z>5CBcDLlCDBN73;2KhYXd7eX8>XYGUG=0y0)`_}&DGzOz?y)CJvz7I1h2=C|D3Z|b zEt=&Fg+oy&-)?sM8$BtMWTyI`)S^{mC~mBN90$hQ8)?@0BmuP7&~EgXCl{ZgYl!uYWPX#qE9V z%$SZoS~=x6m<`-(AUoeIm|y(ARIq<%umFyX!A#T}btM zK|Uq>&{@4TWIZC_EnydY2>AU}lD5dXSWw%g0CxQ#=S8mocWl|E`T4$ni4 z_;o~Y8qGn56vayjE`+XCk7LV!ZQVFM*BXq9(q@~tinw40B*9p;DW^KU{n(T7$}8W(uGE6!EV9| zd6SF~Y}1wWYV123a-Vn^kz8rvlIR4?&Ymf^ZgoLR&O3u(<2hwL4{!C}P{PdXzm@lb z(g6YK^w8bO|9tD9_W1Y9tk_vCKxJzcu&3EajmAv!z(iLSBX7ksEnHx-h|`&}dY--7 z`FfGs*S%-;isC{hkBlIUH1TRW$y)8I6dx+}jZt39Mb;16==eMW>VJpcc9A2yw;tdc zIw*mM4!Vqp6H6?G8CWk-3ufEH9+Bc+-z;;J(|y!^!r8EKYAh%j-Bf>6t9U$F7-%(e z)BT+7C6R;?Ve4O8XN7fLHM@@pTey#5kv>4ZKy`qFP}mSB?G*Zjc!a@CeX2+O$vR(j z<>76F6u+W4z~%7D&SE41`I*uyaqAakG+FR+K$dn(p{M7w8yCTYCu`7~IAe=(m5=@r z8P^Cbyz7U_xnOsYpl}%KQU|_=7L4|3y~Sih_VQt1@a}&{{!s~X2nm%~J z#Ofl0MB+i4@?Ja>6{_;%LMoJ~=(M7}qqj>lKT~2^5UcLid!4hM4IH{G&zHopciNUN zW*49HT%$uO0Mm!yX@%3pBfQA??N&Xl)DOI3a$>C44fUkzT2#Xy_pnW&@d$W_!6+#d zq!FvgD`2RwKLAk>O2)Hx1A@qe&3(GxTOGJ}TsBx#-IvdLE2Q&{Sznxi=<_nnse^_w zD7=%riL@k@O!N<001oVXC;$d^1KE$Zd{2!jTtcmBBE8n1iiapgSa@cYA!50#vP! z&`^S2cG{^tQu|>bEmvJ@C>wwEdlgh&6i$HuAgh64yGVUH%Y#^b_+I)$;N3kMA)MiY z3C`1u8)}=>t+vKn@MNK*aeHGwInLnerJ?v>2d;j}Ja9#L$Pg6dI^qEdiUyrv|fQKO-3GL1+7-?7n6rDZaIP zi;_zIyL^HE<~M6FqW~GVz}I|xv}u@5d$$;Va=)Z+R_JKtDzatWHY_y&kQ-5O7hg|7 zKg|)v)}jM?Ryoz)_GS2S`WDLiPBrqf{N9*?DrFg_RWB=6IJ)1%1TwBqO^-26%=F47 zAI-g+@uhwTkIn_|GJoE_Ijy=xFA972uIBH1X{Zgw)j@mahD-9qjQxEI<1eV3>Z7>o zAzuzYTdn*@uonO|U)e|>kFfi%IOQo8n^|;nZA;fn5yl)ms3zXP;;d(FW(d@Tv4B6N z{1Vo7fv%%e`?fY4I?)2HMTA93ZW{^A$Q zT#>E!`u6F^v8m_rW>*fvJdsyxk<}l2#-sfk&Ss%@@(^`M&SFsHG&iUW`s5JZ!>|b^ zq|rW(o7>zxk9;y8x#MMY7qTZ3jZfuUyfT-e8lx2=@?pQ`q3`>ksvcpVU-|Id4zL8c zfEjt?#iRSY1WVic?8gO{-xYgwUo5(5D%I*foLFO}f%#v~usCS`Y+|4_+=|M7(J1j5 z$#I#`FM`-#?Aw#f$aT3r&uM!i>b(B!$GavLg$BO2ghHXbq+^-{l4FSNnOFwvj%7%pJ}<-AaF+B^>JDuwhI#8Q2iFt%!2la+rHo*Wdvm>3h&`) zP&w%tf6<>uZeBr*4`K7ac3l4mJ862g79jl|qu|)9$J>$TiMI{*6Va-2)T?jtHW&4B zA$vC40-iF@j?DMo3^F}8dI2qL z>7y=Qw9Dh_X4uVAwD&~&ja%It67kU<(O$c8`flRr-#5|Y1SuKed*!F}s9zD1uM1A`Cz@brt{-8t;%icS#UcSROhi!BcU4iziqyv=as?Kj}Z;*XAxFYO${{*wkb`m@hJNznOaSqT!Dz!zE* zsuL`+%B2hu3?Fyk^){{-#Fx7vp_~IoK`USM{fa=Ah5D=CX&wpQu$qCxyI%n=lmCtA zH7#0C#6V2KA;u7>Jzuz{G4tY`?)O(}hG|c5W;Ytjnj~$zH4}{vZnIY4#=gw~%MP=c z4{u*excDe{eqlkDSUd+_NPt?a`6hmitK@-0e}>G$Ew5(S#gAKPs6oFJSkFu4M^p3B zP){c#b+#)dR8%v0@72R0dc*)L0iP|H2iGH@8e416S~^B!WhDWt)#0@tBP&RicHb+2 z%J^&H|s=2oSj1`y(~i1 zMil9bZ#yMRO!j{8Al`F=P0|Q=2l$J(-9rqlmpm2D_}Rlbg&JvG*sfMg{2SRL0ty;5 z@(*{V%v$V_EmBMq5rsM?01SYJo2tC0%IKU{P4@I_<9wRmMzNrLXHGs~+^PQnvwo$K zSxfm0(!S^8TSN<9$C@)gKw zMnsJI{F@@ui%y+i@%vuShU*zgwDBOa%qTj-6b>_TO7(Zzv8C5Lv-jscyf$k8vJ0@h z{3U^h)NAeibg6=Bl}SjrMisRiQcZE&k|tC6gdfp!I!$o8``%8A!ss5MD{_oJ08z0} z#>$K3%+FlGhs_Y63SVJWl{R1{z`{%E^7w~06A^l$G|g#{oOL!H92=2%=>-E@grO1u zecX3R2oep|N`2@ncgmK_-#Py3#W!+~lnP+ZlyB}+gzskO^}xywzG68jVAH}mQZS2p z9B*`zOLeVYA4pOtJ`j`6mXl&dxrNC|TW%_v28zuA>t_J;@ZIZg5eJ5#sW1%?NHn;s z!?Rf=*^j4dqDSdtMwF}EYi|-_!!R2QSjKTDVb-HzHIZ{=9rjfVVF%wfeeEJMi7^=i zod4<*+uCa3gA~ZML#S+C5U^Dz9deH%5(yvMA)oI1x*$wMyjEYawk2=~)A6iqN=;Z^&hfbg#-uB}?HS=lMcvCyzg&Q1re4O$y3|eK_=mwzWLEW;#tG5RkL+EUdJm4iA0r zqGNrIO=`Te<&#)|%yHTJ=JqA-d_F54&RnY*`n{Un?&8Z8e6JsQm9G#~2w5oZ?s`dk zsi=c)s5-JLgLcy=ta$?a3-+b&2TJe*1(yf$9TMY?&j|qNsCvmSq?(dwqxv5DJLWvj zw^2MG-?)t_7`Y&Xj=0W44BDlF#Dxvjqb<0d*_GOJ$;4ae2@&FbMtF3@x-F

I;l6 zp=EGS7qylAJDABi%N6g7@Y-Ano=CwFTEipI*SS9PNq#H(kwCf!{|N=Q15anXS&w3# z4zooP$o#6}77PKH9!1+Xp6mfVJ5K8OgVYUTURj)ghLqnjr;!Bu@UVUtH*Fz^YvZ@- zO;wIdC;QUx$Q^rSQKcM6XL#ClQ-lBfVaGE zDYP64Uu^_JU}{x;7c`}F`D^~w`+gUk+Sa(-Qk4jGFlLSkR+ zH{KNx-51m*O2n*x8w%$nvF)+Mh)>w|I&MK7uM&U<4M}&Gy{lNWykIHV$ya=UWHJ_9?GBmzO-YGitd!JhP8p+2AGf1IFazU$`Izp;Wmj8T0xO1dI3Yy4>Y)!P# z?vhpNjFowg?6v@P^p`t{H~(SNC&{8FHVM2ye3#3*30mEu%ZUFD+Vrm&Cl>y^{{^a3 z7c#Xairk}a@*H2SLIS<^kiMxuxXRtV6Ibyr4=z&<7&gPi!k54CQkXXvofpflw~}^Y zk%pbXe?{IhMjj)u$(zB5K?sS#6?;;+`VO@%QFR!+p4yT_jx?BUHHA2JnmH4v^*R># z^-Uzt{zl=%@O6-Z=m#$~v?P@Qn$(`nox&eN3(lKHm$&2Xv3rlZIC;VSOFUO(k5FnoDmZl!MIm z$E{)s?#`(y4g};obMtywWGSbODu8%UeaI^Wt8b?|$ZBOCLhr+_y%V1ASizWC0ZvNP zJ$qrrY7YP!;{<(r(t8|LevEM60l4J zK;|MIXc4(4cnLl97~#+xZhhMl2|C}YMWDz1^*K3*=S*%aN{qFp0kk@*f0wT||492r zP}nwK;jPkkt$D->a^Gj<$}om)qn0!4~kM(h=?UDb|KO43J%W*qghRxe( z!P=Ugt5aPvcL&0r4ymj7=y7DE2$ZJa!YZ4)M4qcMchdOrpgkw89bz1KjQ!Dt0BBhG z+7FN4gvV;;?p<(*B>3ez7emX4mEW~{Z#z8&+;rRp| z32Z*YM9BhQf4!;5O>yAlZ9LiW$BFSjuLro{UzY**nDuSU4ibgNcG|yI9t`y>w~8FD8qD!LVOdXkKe=i;8|uJcI1Rm6-I^z_HS-3cI8t{e{9I( zBV=Ci<8e-5;9@8Qr{lV0L&JEayJ4-ggTPnGz$Uj$xElPPCA!0xR$ zlqg$q8;dzZODqHdT-RH>#vGI)z0j^={O2k+kWD`IDVQ8qp-D&R%20sX*cw#_! z*zNBMHUsQ~z8L7EG~$CsINPOZ&$tQ$rOJ!FENdi6*Mvy0YR$`;X}UNEE5Y`tT2Rvq z-eE=q!F#+{dIfp?;#0dIkycGxjDasYy&xn)PL{$Wa^2^J&JuV9qc-6|B%R#96OAu# z0K^t*hP1ZPwLl7+9$w^t?Jp;n}>LH$#7OsWbqn@nMhJ> zNXmrwuoCt4_U8go+ErQ9ls z*kXjP%a2KD`<)Erf8ug3DUT<5 zr7qWfN63#KIZdmvw&k}Yv6$ffnopB)xB2~uRX6by9sLv#1nu=~HuEpSn#RT7`(Zph zSW~y6R~FJ%anwwUoZJQb0dBU}R=9>m3)H6|v!Uaxdn3X45m%y4AEi>$3lc0vDOcos zIiE=Znhvl%@D9K}1Y8k40+_hSwrc7A`8=+j2!A*DuYfm85=j{giC9ZbV88lM;A%u^ zcy!t@>IAWg%yPaX1lB+t@AFJS(r}-=U_hSjwn|N4s$M6tmqFg{obo|91lys7JaRK7c~47F}U5N!@i2}H_OC!y_66k6aGn^#@N+g8o%!ww7eS6 zhdCS`AY{D6#>MfZZW(WBK@;jS>eRGSE+aZBaQ`s=l+j z&Ba*M87FJ^ln{mLQMxMCX#3_+zE+k0t0a7*mtp&;l#<`pr;Qof&KH=Y>2Y}Z{fYS!im?6+Y zAQK3jwAnzAKG`9? zodf2|{8m;zf0hlgsD-4%`*#jLr%!uz4t0z&9cOb>t*vHyz`mu|n!0}GwcM0?-i9>S z{X#(Mi;iS_ymap29miveSX>ev$xrKkpLTcl*2qQeh_w=VXw&vqOGSWf09)>hzd{3h ze;ZThikYrFlW`B(@c@>|vu7A|n8>9$%^VvbIp2%gyqwebN{YY@sE9VQpxohU1)|J) zXHqaXJj2~SE&;YFh=8ntyW$V%P@KVF@ucSQ!OZTJH%6KG0-svjWWdtpUzmN1PBX*hz!s-sf>06Q+fU!W;$$Qk6^T}V=jAW&;^b0N0a8!A$o9)^huH(z z`y^dzc{}y@wZDo;k2vNJbirF>3_?&VUdO{;9{Wuxsc#BheTAW#GyE~1BW_mVl%~%D z3kA$rjclb4y>A?-0;)kp#w!yWrIqjJhpK;~#!q+22fkw_JTW+f_%kh$tpc=%WZ|WB&okr7>kp- zpnfhMA=4jH3yf}h{=Tkyn6~!1Dsu<@>;Cyi=F2l6!x_-%!+*+XU9#|!gxE&Dzzq2v zueX3}_7=<5kx*$%Cshw$qlrm%7Y9YPG5uv8)rvsEvvqMfwU?Hp-T z0+&G^Pmbo)2peY5u~M61oDhKUE%F?cJz_)&1VD?AJ%ar+GxeE$J+ce>PXYqujGS?N z6i^u~gjUHO>5+R!b&q;l&exgU`1S!Fwi-TI%!(-jL~=1w4l>4+bl<|h9aD}Yp$r&RK@NQVTt>mY-L={blI6acELd5<_X(8B z^hJ%K!li%wz{FTvga#<8bt+@kB5mj@-w@s)p%55ohh>7PxxX{5xR8C7mn+UgUIT;Z z*`|@DRg=ZJ>aVOHA$=sA+F9i;OT3bWNg<<|q!dIbgHc^Fb_9LwhaYymn);|L;l2J| zIvyR?PW(9QG){(DfW^Z@_(W+ua#WD-=9;_Ndt|-0vtr=Rg1r+NRzMiHrf%*|;TPjx85ZK66?~qA{DLV!qjc6@}^i4F(XSG5@K2WkQ zYU*SFGjhr9JW5)0ywo!7H7`FMAg^9)8%j;)7sjhVk^4nDfQn9g?tfb=SU)>BwFT&b z`0<3SkoI$XZtmrb+kOdk&irrf)-b;jw;M^?3xN?5^ zZ{DUVlLV4I6xueZRsvT0)1AK5x;AzkiI5Nk65~ZOJa#4?mXbvhFGRt1bTcndaz1X1 zrcM{5#v>nlm&S+{=7MUo&U8;fSvFm+>b}K+hKj|c9-CljTxrbvO6qVmbXJzY2=r47 zsgZW6FV0WeciHpvtnIE!nCqlU(b=@w%BTj>x}kr{UFP_VG7WKFY)AF7s5%Ndou?;_3g#pghLa?HNF>I zRjt#j)Pjh>TZ5g!ZHS639dGm=BUcMmOhysy5TQ;UxG39vf;~b{6ryD2qwHOb&IxU7 zCSZ9JtcH$9qJY%~kpRQOi#xL<`rbS^2_G!sKVr?RVWAj=@-5wcH5IW^fIqH%71Cxz zVM#~`0uqKg0K;nJxyVToGEV_l!nn!zb3c_A;{I~kh&mfpAi1V?caD%fi>DKJS z@=4%bFVv|@HR~^I-YGGwK7%Bss=YwrbM^Lv2Z@T}WQ`(3OfoZj;q$PfSuB zKcxfX!4!TRZbo7n))AEiBR$JZl9V9p`eqpgrUi?xL5BzV#Hanx=G+A{Xl%^T5#3(M z)p)nBz2yuJbw0TX(rf|^rYfPxIc_Juz!`{ao3NV8Sk zU&75kPuIr}z=eVgzU4J@2pIZl*#S}ETYT>Eyrmqzi2ZHG_u|^J?dJdqH0{kXC(7f@ zBBnF~VvPfwk35-aK%Q`!0W(9CkWTt>tv3~)+gYNY4?(L=tFa1R>JZkl;vrwA&y)qA z$n@#B`~+bQFIC!vnsc{{Tu4#M3f>Rk`Y|G`mp=0MhGy0CBq}W-AQKhDC7|8Gx>?FrVT-$N8tH@G8||{9 zRs7P?AB3Dsm?};e4P6!{0qiSfD(p8`gM)yOUnPVewU52x3|WdKVHt0G=(|{Z#o`}c znpl(cRmxot^w!&RenG+}r1G+MIqin`V ziYEe<)~<9Lq>!dsd_0^Uf~lQ>V}YX|uPTKf>{}59eCZO?H9CC%(6mdFZE@kruulG4 zGB^IWoYg4JrPe5BrJsa)-y3YgwU&Ao5MMr6lp^^@muum}un)&<~*i6aPg zns;isVeY|bSw9~;)|9q)5TU|HY;CK-3=w=P%W$$eSb^O{z{D%kWLGFbI$QlP@$twN4{;SQG;LnnUKu!APE(nY<;5SB0QEAQblS=QinV7;6jv< zKLI*02(zim;5s>JE2;5l>EZv}jmyu-U$Nibjnf~LhVUltWT9$${FeRw9v$-ZaNBZf zyByjsIPvtM>JAHncDXIUn|Yiir<0dJBdH4jcYW>)*7)NrZ-N>FGqp*yY3SkhhEFYg zOTNezScRDi(v@0w_JQ{OE)9gi4_CN!Sdby;z%z%aPYS-hWvW;{=nd#aj~cZ>>WF&j z43Vrgm{^%=Ay61+2%mrNhwHbUiu1iVGy2+y!QR_LXIO_I3j!eKl>>LG6ASE|iF@BR z8Gg2dC_``pO#zK>eu0ESh*75q<+echMfb?v!6^cXAARh}o=Xg{&nJkmQ>mmB&E}KSe*fT+qDK%)0p5`Fgv%g1ue)f$)gttdwZtY83yw+Inbo zU~+lBqWO?bH7~DtZLYgWri@qcw>%u)v#bh&_d|z1;obf zw4L`WOg22WWe>(+6zxx~udpC}B8bSMdKOr&Rk0rJ`%!mI*B&293_+61GdjIO@k|{m z>6n1ziN}ffBZ%3*T1>qvD>2z61C~zS-HSS%xbTUp$iQG;6XD*%1xaIUl*)M0*Yqs} zY&fYhsorp!tD}=;SBVv}!hrnD7XA1IJf~qmF8nM~i~p)H=ZcODPJqm%`T;5h91e`g z9Ksx1ySIEXHS|u5tDfE85y_KXge~4O0R6~*hi_Tdqarw<0nT=1Y`?oJnT|-na>)l# z`V&eMGNVIm0~h_ii#dl@UMCB2HK{@Nd$hm!a#~DcM=0HK;-xgjlAAxG#yxJJcprZE zK3knr^Ft0^xle3&AsZulLJRizL|jJ^y|7v`Th(we8n`%$xe!a$sNSH4#@KLZYZYB( z7>HI?!@xy^pTdlHbHBpAhPfnuE{<{dWmlnoyAPD9v*cBQ#R5#?2n=}dt{eVtF1h{? z4L;?2`F7x~rCWaidl>KnKHHew@mHEZ9Tvs;?ZC^T%=M@Y}MJdsp6>N+-@ z2n>w0@f~u-{$|v#cYXV$0=zC#9?~^s&fgeOAoB@Z;N;s!Q*cCdl}7^+RD55kj<4?i zCFi$C)csg!+~l;pG3eg0CuKlGI=m~07~$2M0KB=qHR|6jbG0g9#@o|FH1mOe;DL{t8w-ul6T(PpCu?ZSY4irj*61k6lc)9Q0O zy(c37+Y=mFznIsgf7?aG8<_uT8QxaOl!xQ68TL?BK6n`&Y*{Jn5QBmD4S>8Iz4$c2 z7D+NX=!JBIukgN5hI}|&#(xjm(>Sl zTSp>&UbuyOM3~Jgh8;%uHAwv3R_nj8k$lx2b8N0Sp%Qld88i&yc2f=TaDd|HX~N`PxZ3cx@Mfd0R@a$b?`=ce z=jJ|aq^;uMMY>$}dV~XRPhOts`*(2TIyeaS6&!!)zu08jB%foBcBZ?U8jp$CMDJyy zY~F{<2gr$YtEzR{UPEb3Y(&yfpEKI-Li}$hJR3Tk#22~5`TB3W<@SgVe7|W(ChUj* zb<=4qUbluA1TH5azHFuX-=}~7`@IwYZvM3w#P$*f%e`v0!Z%=HncKA-r9d|$X^-Exc!13r?A4q3y*K4oIAV*9QP+?>OZe6-=Br}&#x-a zihu6?|MwU5Up{&2pYQ6we(L|@wkdCuxq{5(zaBmRXWzqr{|)@t186^D?tO*B?>`S0 zYR?|pzsvt|xDZ%A|LZsY?mttS|8uYZdtdth;wj-y_z&uT@vL()75V)C;937U;QsTx z`0qcQ|L+gaJuQ3xl>WQJ^Iayx{AVOssWnydbWpM&f<_Z@zr z-G9I&X;b_ixoqF|6MmlM&*TF&?*DuQy4~$JCvNTs+?u?HJl)UbXSMD}J1(wxLZpuT z>(>3)|K}xeSu!X8cEaRoe?{1{3^^_g`?-7D3ob-cAYnUJ?1E-!i>Do$k!gCrcH&RL zNI>QamT%v^x-SADD*Vph(qsF5fVQC3{35fA6Nk}R=J-wf@ki~~5V*}oQr0^OZ3;=c z1&ZRgEPxw+yb-S5aTXEj z;6cc)yy+Ni{lEDi9{ozwhp) z4@(V{4u8K1GYx)|#mCR#GvJX9!qS$^u@;P#2rZIb9?m4eMv=Gjo60nav@o^0E*b9l zkyiO2OM+`){O+(q3FrGP6U|_5-qHB&uhOE$1z?s!|NDOb4=)9(5ZZ!ezt~?gQhEtk zzu3R$hi#NDJPu@{f4zDO?``89u-zHP)+;YSAcgJVwTP<)o(qG$|JC+t(%s+E5t+C6 zL}VDiEvPS{4iZXlv9xM=L3=~@{SjL`jDM6vs1+&UG^h1gQw;#)6kxt zxhGq^RdHUH!;0jkb7rL0=hq1dUvNnEw&(F}=$bWVr_nvlX?9AW0hp%>68>_CkLs+; z5G`e?x)`u{F!hVk1b?9^QT7qU(uANbM^u@e{l}9%am*>cM3>OH2(GunJS<0kc1;f$ zOZUsf)<@+#-Vsngnwf#^{!v~|ga1wsW-jSl0WcU5`R0)yJ{D(@JqTSB(iVRy-pfJk zZzxAP;j;zWncLQ15h$1#k*F;+X<5!?hM_1E->?{bG!);J5J!zQorpg!0)V(a6cG^V zKc0_JQkXQhyV$26IUWa_?bDIBrB;q}0iY(=2R4v={H~J&X+zopOI&>Vqt_qBfi~{a zwtDSnL&(E!Tcmw3(r#CD@Ps&HiJnHpmFwz>|8&rM@ZVtj5%Dl+ejrzV$aRr3e2nAU zf;=sFY6q#g|B6OF7v{m|)azF9U|IdlKq@HI?i;0u6kGDXJAk3k+FA_eX42l-n;t~G z8rk4cU=I!@q+C zS@kzV0QOH$+qaGWf(d@o59wGc(rAK)Brr^v0da!@!TlW~JrTZB8Kut|GTus#GQb}c zmxb72o|<=3Pvh_4GN2Ah(P|Rb$d6rw)93W~8HOfYF)NI=0Wa3Gz)?iY0-T?WK`866 zcR{=(+MB8A6&)~W3TMJSxW8a#LaZGk&xjQhCBffxpa30(u0%}TFgLJOq}S>iE-cLL zXR*h_`Mpc2WTA_!(t&Na6=NLwn`%DlNbDca5wy$wHMlD_$(`=I#iOhvgdQFa;^hc6 z3mL{4SPrp#e)cQ8 zu2K054*BE6o&7~Gt(r#SebIu~^@-@%xFZFg)Rm6M%NW6Mx}i6^Db`ph(f5bcL_G3< zvE6T>KO%@k&4}Miv=c2t%tdT=LyQchX8Lly>|gCTHzVLz_?-?@(#Bz}UcP85jw{sY z`;IR=lX7K7Ri8u^LI@w<2gn{ia&}aIgzF5+?^TS{(&e4!YgQc`H*&D^#5HB)cJaWr z(QAA#S@W}hOCqu|AkvJ!BY7Oi!|l0WhV;=De>_w^T@6kVcSu?kF-*?9@%8xGTEO_V zD_$A!D_&D^c$@s-5M>`Cv=?eisJpO$Li(0ZJ&~#4jLH>S(C`*K@>loFn|rc%R`V4X zzRTZc`>?zewQ{d}2oiJph763}LiSdx*urqy2hsIhp{=2FfBuZH(4FV_Id33IL1;5N z-GcuA?{`XH^oO-@632*1Hro*Igl0fqYNnrcpkbVVB5|3h?0#A9VhP8R7c^^nJ@C~| ze5U_*lFn;eQ7sChza%0kk7SS}8GJ*OpyVX-^wb!beHF9CIL^aYu z9G_@lV*cp$mRr&7^FUw26nRXsWYEPvB*Gc@7ldJGtJsjr`TfBV%9rfR)%4DGq&%5@xsdqs5B^$5{2|Tl*sgyx@L_vMK5A9zbfM56|8xSuG+2hX zK?Hdc@Q`y67j%wDm!G`h)Z5uk8i>C^^{e&Ls)~A`{44XkI67=c$TAHpjB=Pqgl;2l zMEjlEKm-~;PoQ>Z1Cp6M*dV1lr}a2KO7EdVM{vP%_4+{(JE0oqUF2J+%`pk$T#k|3 z8c~FW=O#T(rBk_Uwus@MrT?}+p||hEFCStY^x1?OvOLZiO1jm|?pTO#AhA$XebYdR zNSM{p_w>Xt_kVbgO}r7fQHn+YkGac=y|6nEiszkEVC=s+aLTJ>%Pe?Hiua`+#GC&x zQtSDaBw4;N5>hEY$pal-wl>td_$IQlADEH)gv)rqwfU^z6#j6Rj%DR#KLkSo6UF?Y z4Bmoacd@H+O-$p3I<|ztA22X*58m}nRoGANFn-vNbGSTy7evO!;3cU@mJ@-4jP?us zk=9?m-t7sv1h7bfbb#A89tUPAVljqh3o6ec23O{&zk+9+^KQgqS=>18*6j|8Nz1m_8HadM5`D+jZB_4W@eb~oU&$F2qx)j~*_`3_GZG{h^dl{Yca5}ZWA5c#mhEZVXe zroM3K&j41A$Mflu+JsF2Msz!`U>iLCr^)Y8$LJlcx?qOM44n zSIs`OzRV}&^ZF#Q20$2R2$J?;(V!PSH2-_wIrI*EXo7Y|YR5>NKTq!jlD7RRa=0NT zd*=3$5o_6(8hpWSf$YZC?2r&HIe}^2Y6wlUsY6%_JRZ_Fi9envyloUn4SwHs}){lv*Fg+8uk9&X3(PBNBLQi zn?WIJwEPg)sp-hyz&4Otw(^_4?lyu2Z zvHZ->Jplbu#q-cA3J9_F@nz1*sVSkRDw%qk4?Pc`v_x z|KYW=Mh@zc_JlW94v0xUJMs$Ofvbo!2@Ocs(9?Txrk}3F&MZV}1KyUxPv{=9+O zXNm&+6Jug=d!9A_d6~iLI${ODPMDUE8wudB*ZdM4Q}(zbus8T&p%e!s&ot(7uy;Nl z+WZCj!k+AZS?=Jydp$hYp&6X0{Cn8V` zWI1G?ZNIuQ(Fy4?jV@>nmfw|Wtl0<{finf%JJ^Kpnz0&J$>g9E1?p2cG(2>vrB z;>lU)Fa4vJ`qc&WY?wVo_!?L++d<`E!{&K{1yz6V)3>tt{62*tXfOe?kI4i}c!5}; zA5R5{=*uf!KPUtpzq&kx{9Q3#w{ZQgHGAEGwA^gPV0%L8JC&0E%z__5QZYTC%s;%2 z7J9@axvr=+k!RW)82jGWnNfk5ZWYD}%#b$Hh?IIvR8|&Hf%D=A*>52*N=CL9r@88Yo1OCH{I-*9 z%lKn;uj{s*?Gw_0RjB4%H09+Yu#Im=9OfGjhH_D0OU0BI_8gSkFjDm-bW@W0VT9(? zhy}0PW%v5E$i}=+%5fRW>t2K#x~~DE3W_ME&j&H?Hsir1K0- zmz?A0W<_nUY7KUko}T4Z-YWw2v+Aveu(@A%I9T!gcc4EU5A;5{J%9=QbrB3?GV~iM zYl9$&E3gAC042uz;8%D3mCeh(r$wmet^glva^N9w8{F~lV2`0vEe?gdp#s7Qb3pR& z_tgh=KRM{ob94`2(huGGVH%5)G}avq!ln)_v7e|Y_**BC7lFic-f^F@(|kU|>|Kl- z(j}fXU?XFjUXPfYB5*k>0P z_5-+pDvT6H)x?s$#_Oo&`^fnT#{{^%skuGLxeFcf5xZMK|h7#Hp>vwXoLU9v6+=c>xTg2B9O?k;be zPXcQOu$MztawYG}?ieP$D_IGE^JhbUG1&rqNcZd5yBmb4x~1s2#W8^<8)EsG&xziy zuVQM2F~u1S#SaqDA#U0R{G@rYLga8sAJ{nhC0h4gFBAM;nA_Py&Ku5EpwU>Y|3OGL z-yGd8sIHtZS&#i(mo4XLlwe@$o`+9(9Zs{x;zRJ4@f;ux5mKPTLS=-1n!aIJT)8u1 zlTGIgADv{tKR@)l_sq)Ygr3fqW|cvzRCmLiwcu=ByNDYgM6L#vpCXvc9FU(LiZHJI z%&kwS%d?NiUOzhMAk53Jl>|lMg!O(w6%X9cib_Yh2YPfrYr2)atl7GqP_{56^ zf25(#SN?Otm{04qtic=|TXKka%L-8{Yv4$d`>y?jst_HF-;bGkA=sXRWHySe%G8k+ zyPpb{*HJ%!<8GDG_yy7N#dv9webTRUa)X*elHg-oANd-ODkmq@c}}(?bIx2t%fe>- zz@?Z3U8%dP)=W{FVDab7G4xRRnZ9`s-&PNPAD4WQ8ZaTK1#Xa>FB-jgd;P20pO|;5 z&I&%v4aYz7{?~wsw;Z_TT!#22&`!G`^z;z5@WX-3Wb+K2(wFV`#dod2h|B#`gHwB$ z2G54P#*IxrkY*jgnxb9fBL?wkI8Dq2iq}9Ya5^tmG#=#Ru|Zp|<*Qdo7m2@?whgGv zwGD;jB@-|nuR4FKD-g`w64}~3v)U}F?^~PP62vvG%kW`n#knTc<;M}3^upRHQgLd= zsEF@@dE6C7D5ud_bv)EDT-ZIUpLuf#uXw{}-I+NhhbM&`M*QIWSzN+$O9L&&zwU(h zfoUw#pa}trOd2`Ja(spfgIBBXheC*;=@TtUJenc6F6Xti7Q9bOY4KTmEBev}LcZ!2 z!!ck}zVNkLWI(@War^OWkpIKzO`^J9jzsokb zzcOxlt>*;wstyRd?jGPWEFl=`Y5=JZ?+4ZAWA0TZ5hNf!y|}Lmf?y~iEC4Ye?>os5 zv^6@ zLoPnQ8hkqCYy5CTsIk5}OLqHyg1d(5p+YmN?4af|^9PHlnHxFz_+|k+;7_OxU8}6P zDpyK8GTHMCs-UM|{Vx^(n9QkLpY5fYZ2QG~_hh@9;LC*4>+=n*e4>_>t}fXR7BIY_!?BCsTOg;~zOB&&0oFXc?C3lIxe5=>xmVtgEm?3XFAs`} z=tKp*qXg{=KEc=AFB*tS)?6RP;NduhLv#?a_gYyv(2pu-8Y@sFeH@S? z6uSikygZ?3kc6+glv-M}_DADGVf3;MESE@niQSz$dlaBoSG=}w`jxLwa8ZDUDy(X? z&6Z+3pyI}xn^!MeQuW(?w;cdul1Qy_*Z6PB=GF!^%Oh(tVW&=1Dm2NXO@LoLH#vUR zYWc7>!|J;Cw~+cIV-OT=0)0860NO8q4VM>O^56Vkh`(9SO>eKbab>EB?T7(hJ|AT- zO)MTgy*4BlA`}vi$2dQ`a=(eR-|Gk5O#4j^@+3*=)jgHzR;c<+U;^;}KDumAC*Sr` zHo<5UIQTXa^MT{?+^$s`9IP zuEtfwj48_$Atr2ix&;Z_%#gqgm=rkqNSAgL_S5~01`0yg#uX-j_%`y16~!q+Q^Pf` zb3&!!2BR0BIBx`Ct1YWayt}3XeX{wyJIYrB6`7y@+8B7$qtfgmv zII1vC^ObV)NmW4n?d=$qYjvdKBSj!)j`7!Cy`` zsw&u6e-#+`CWwu-N(wlr(1OR$AwW-?0udmb3Y`K2eL9GXQ}n6D*AiRRy#PxCeTvtP zj5@bjn!>T4tYGi`fR*hXt2D*!ZYTSB19?3i6+V84RRi|_8jkIL9_$nb2JTZ2sP+S_ z(FbI?bX@UJP@%vCA8PwOhzm(OZKP@N1BgH1mfA9W0TlhJafc*bKopNHqh~b)WFdS# zzMSc$Qr0?%=eRQA;#z92^06(L?NPoGh0Q78U8opG`-7?1B=>NFA6SGOc?@jOo_PNJxqv|aAArM*CeRlz`PbOazzM%jSlX=(gft4PRTXjAAd0dv2 zuw?y|8Nc)(!r$vIdM#AHmfSD8{Rpc#A8o@~?HYqs`IM6lqTjudKT2kIXh&4Y&_}KZ z){OAQ=6F2&F6Fm51t$aF+`qSNK*E-|7eeafoQYiMHdvZoA>gblxRAFk2r}}olKTzF zZK&e!?<09do`@eryLo))cOQlX%f@W83B80MA;3&~@+Nq(F9tQFpXbFnlp6=sn~-2^ zhC}J5^@eiB=z+PNFO?zXUs#5^_VlNM9^$w{fFYSDWpOy5BIacMQe8d&QnaxJkn42z z!;h?bO)<-E`};zsX~Nb>avbmlShymt869ccX zaa=-EkHDdW5^sx!-javV2!b8{HuuLV=aHDz3ox+dK)``|9T#Pb?()xYx5m!U zrI-FHtp;~@_xxdMhXxW9hoV?uNc&tpItc3CS9fH?Ua#pG$ldj((!_~%nNXcALC9$(vQZeY2CIfYInVx3~dSx{9#!!y- zy2Jm7cL%>jd?^y}RCD6WihuVT=kd$SSGu$zpCIGgx#NNqUfwg3yv^h#Akr3hZ_o_X|=np!c_L%Ona$Bm4~y{EM>JVn)V(MFgct zFCEdS9$>Y`XkZ_Zt{0TpoZ=&&^PYT~qvwv1g+M13Ec?TVZ&mQ%BEX`9nLm94XHu2c ze$g&X(%hsX)l^ASg+~X=!v6lDS_%`8ZN3M`tsv?Az-)IVv-1?zT39>g-pHB#A227M zjnFNae%^8FHoQ@@aLxnG+&>-#Vx2F-CVj}U(Xd`t;cr$&yK;E0QWo<3>ICkL zy(F+Aom->*ZA8=mfzh&nw_9Yx*EzF`&0D_$i(k4|g?~Q`&%+@6`3fLo+6T4wQtq$E zk3qOf+_mSuqO*Fcw6Yx^~yQ^gE#sY z;T=ytJOwnez+Gh0$JP>k~0*Rx-gd;@Wv5TK%zwS*DI;)7v5j01;oE51^H^F zisrQ4p!`%KetFkpB0W#`Ky;#ljF6jhMgXo_8ts=@YIPZEKr>OzaW^u)*dj#wDsKL{ z9M%y3AJyimy^YUeSF~?g65=?A77l~pF@sxhW>V`{dF&0^hTnM~50c*#jZ5@%xTc?) z$%X>Z^N{T|LVbL{9d@5*Z`FV9y4#H(?1uUnzKB(agE<_D<-u7Nk@N&%&Z)?8%Ck3e z)#gzZVWlL}*D-$4E6OQ}URqdQMBN)t!lEfZ+DM!BjnnT*d>@|GWPaG+{mGg!>h>8Y5oZ%*;2%yD~} zZl9r2TDJ(g6*TvGC;ZZuB8#j}L9*0Y9s2zF+o10G<9&&%pSth!PsggjIy}RFhw!ty z0AWbgaPGFP2Sk5q+~wZ%OQnhcYn)x#?aKhrDIV z76DqvC?}_pNzID@A^IiOp8gK@+(s(pZU1!eBBJ4Z_lX{E~okp99Ixl0OF!4SJg>V z3>l_9t$s^XPe?moVG+i=?>W-eCBf&7LRLa5Ht$0~Al0vROp?p4QISfT`qmOUID8Q3 zK1dpGa%MZ>3gC-}5UBgUutvWzlGXClpAmsE+8mW-jR#$zcbC07g^k&z_S36!Snp?3 zyL;v{)>#(Ah;45Vgt4sNQKPbA{l&)zM}gm7PUt*^-n-w-qp+`?`w5#PE!}<`%>pe( zxD~r`96=aPUh`j%f^ z_SX1FJp(%cKRdq2s^7?|Hji>o_*V>(!pDXi_=Ea@fCjc?dXf8CaD+*(*XM279x_ck zcrZ^T3nyjGl5h76PmNAh0KD8|QEv|pf&=<1IybPzFS2EC=9f+L1kEjMP05~AmL zvn$4c&G`DtTZ%$rL?I~470r|rh}jAAxQv)7&mj0)=6h;wQ1Q5w;;tJ?DF}KTV}pxM zv_TzWgx_~1iO0PR*&~yhoM}v8GT|4%*`oE|75merQf zJ|u%6y<}tj`ON8xjov$$3ZhlMR`pW*{NQ8erX!>I^1Kc9$!u__YcDo`Mfff5g4T=` zT|O>ij8sp6RBN}A$%lB)Qz7=D+gT-`1jrJ>QqdRdsW7aE;N_n z+_RFeJe3>1bbchE2*;Y5U+v@*>=-*Kt5e%$To^l!o@xkUvc1ncDnYwkPTjh*0qfl|JTy6M{fIz5-sR? zUn_W#pyLGKCiqQ{c^drvaqhbW{?lC+l#%Hwn0}qhkja4-SHhtC5fzP9G|;YYXg`W} z2k7A!RJWy#ZeDXb5Z!V0k!5?IRbbm)G&FHzg}Y_}{{MVDQvqbp2ps7+=ilhf8zPc6 zOR2gL)7R-D0CK%9fAX%s=AWuO%R3Ah^&rCSH6!Nf)M|6btZpzc;<#_i)Po%^leZ9M$ z@2L4WlKG{%3J=&kIylus)fNf$P~=CYUVNC*Yd;`cc?Q|&bcFJgm1}1{gWuw9AorDL8PR%aGehPb-t_V+PIhV}WRN zLX$Q%RtP>A!`rQt1MkX8>O#YbC;9DUVI_e|Lp4A@$^+I)iVQ)&y<#(9SHpR0a_vzgM zhS$^Rx-JDn8Fg~tRG$Q-O6Aw7#{_O~P?yLM%`mT}--`VjP;!g(?K0Qat4onyj?!Ds zyVT#84`qfN^Llgk`arHy!(<7Q!HEdW>L34p?Bx}{1GW17o;!a)H`=RM#N^&m{J@z?%9J$8xYwNVU?M^5!c&}KlI=Xe2QKXl`Mxw!sld)Yvt6DrZjmgtXooO`hY zL?c{~7}dKvfkJL*Dr}t`@%~l~o6sZ`;+9TEa{m0ee0K4X@Z(k$}Hyyb_8DjCC#XD`a| z_kQ@ghn4wskY|P_^vlrPnH)lF%kChZuyINwZ-)gf*w^L>4jLWD-VsXC@TW`H%~%WN zcf4h?#XTEPEek%dRiH>rpy!{%fiV))aCT5{Y|-a?$U8$GRZONj4PX(Zviq}eDeLDp zD|9jyqS3&*KTCJ@klje}>ZBDmt>>Mub_IW7^#1%`i{7RFWa1^; zYhOK(VaRS*4e)|*&pe&%0dYV}K+rjd|4#zpevh<*m6-dzB1H670ip}`uL47I{HfL! zw)i-UMXv7)6}~RR2(3cIv4I){ls;VHD|1xG7&;GWLsw5_U1Uc5q z@r!EuZ!?}zF5CTA5L?@J!+w7?xYU4o+c`WGZ~{!!W}d=AF1I&1ubHTapY(*aSHihz z1=IqH76^~J;7;$%+3_07^NEwM&i(t)Uv-P|M`1Ij9mcGW4C8hc5E+VFdMTWaD2Bwk z{j%4r`YMz{2o>Dx6&`pr_0D{=;XSzmnT*cLRG!*_rC!lyctw?`*{a})89k0@a2GEpYUp#107PXJIo!MD)Yx;Zk?UE z)au1OQyjxnTcffso%g7It^@+g`86~2LwduadF_Gc$tLvHDmz}0`feWg!h7auv?-wK zxL#bB-#5hS_Nw_zK=h%=yG{kAz0V%u5c?$qM2!9;k$_EJK5*zK^GR+^Ztzxy##+lX zz?+>=Z*Wko|AHnxrmqZ*VNUJp2uaN)LP7GWZg19PH5U_ChL70Bbiu=r_GZnDIHbYJ z=ub%HLow_O6Q{lf_4R_K>3lj3fSLn~h>+@(gxuwIdl%A%C>{B!Zi|Z|mZL9Ah^P7+ zi+J;2lnJ%NBa^P;TN+_TT3>+}G{>vYD^3`-2~{ zLW71f*C`_HX)4DApmMBpXa_lZ+ftCcAm!$T8BXylXA{z1#r)~j`%GWXF_>~7G1^cZ z&_O~j9bglZR$-xMe-~zf)&F|X4g0eSwGD>z$N=J&oGwK(=&{nw|i>`?4rCbOWTL zY4+v&d^_IHRVT7X@LYa9G z4b zVSzpwfmH4NwDm(n9H*D`{afw%Iow55E?588qx&J<@H$AKTX{&N&IM=P_Vl9dm2(G9 zI3I4yIZnOmibk{p*~fr4l4D+>8rDR+H|*dloB!JFC_-SL>m}E_bOd{#&j);A>QAox z88Ypw1Hk>xVodb>f_hqZRl0~XrNhDL8C-Dn;U}qQXRFR#TWRo<;!rb=@?Yb7MBe(e zuSE309NiwS62wd^H%QSDS;1o$BU8Kz`fM_jD>7PwEh9~ z*Es8*E5u&fFUU$a9PX|2m-dd!P~Jk^XRT{xuTF*&^Ux1L?E}=gBI@^VcRx>$!{xOs zji|q^C>oQFPs_td_w%@`A|-UJJ5Ba`coDKAI_ka8vkC-Kc;MMEMN#+b#f5Byw^zaJ zkSM1QaZKIE9GQ{1orglO&G_-ST%NY_nqE*<9CvjbM?su+8eyYEuPT;vAd8vHX!e+$ zvr|akcw;48X5WZ#7Wh5n)6kWYO4CkIgcMA|$&?a$xL=Xh)XHe;5^}aokETUzHL|_i1a0=;D?Y4vRl5*KS&J8XK0E{jDte>F# zG&Qy=993%Y;QDhGo)4?l-;Dv@X)w7uwZ#=GDIHuUk7MBhbz`4rP*L{7lSCkwTg5ry zBZ80g^TvHU&&cmhbcu%xpNhac$J~>7^00$!@HCnW21oXu`kudK1A5*%@9W$d zM%)~OEAMb0bvY0a3cGw=UJfC$AVDvWW<1 z>5{hqDU5rtuNm{JuRspny#0;bpUvs|12YY;7&|q8e4Dwc&r7Os%e{QV@s5aCIFKc& z23j)0_JQOppN=Q@YnV>1b5;+8*uE8a6rG%vhG%1JIG_9Hy$ESC?<-x^pO$YSfEjDp zZP)>p+<{i>lNVI%aQ6Ydh}aFb2V!CKCHD0`TyEQb#6#8+k8!U{XlQ`S#ZB)h*j~8n z`}-9f7c|SF@Gb(ek2CqIooaXgB|B;8`UrTgvEMpQ^JF`u2z|moT7*iXCjQ_#7{;6* zb^d2fWEJtsYekg-lTgjBI`ssKk%!V zVtqGgCX1Z8-5vt*o`LiS5Hq!efw?^mry|%Yf4vaD%#2CT6$jV;MN(u{kkW@RG2zX_ zt<1OQ92LcARPF&2ebV~kiwa{Ffp&P#=k8DY0$dt$O&EKH&)<9n`h~G(v`SjU+7R+n z2fg?-6-*9pt4+yE{mdxo9KTtue9@SfzorE=>nU9KSFSoEECaP2p4FsadFQnTS%Z+T zze}ZmjJrqzMeRL5Q#eVy2@G?P*wu^ol6N=xbQtJH_}D;*yOJtgH*7@a{w%zqs0a>n zVH_h6#_&{#_NX4Bhq7O;G@+>N0)aCdS?v zXS^AIMp2W7dffJYudN}a;uwO*e&8A4_qXr4cwO5Tt;Uh$CIV!KuYLEIrBzq8G91jE6`f#B zdhb3gujvBFnQq(D2b$O^3;!nmbw1^DAo8($^gdxQ_qo6iV13-sgMpI*1d73>-_7m3 zUqsj?xa`VcUR}5}qz{8%@x5{4?S1(Uzp@J;+*U+#S2S$ z%)>vb%%C^!_sB(;c+_mza_@FJL%xBZn&{UJxFi8;c8nT&QpdZ&HTW!Y!Xc5a2t{ z)4L0YQ@&q$qlrT*6}H0Ql7QhC#J5AbEgR>}?{dSd$32uH*tx;1RX0$y_P2;^z?t5Fue>JZenz3D`=k50;G9J+Deya=R^|br_gYm`b8j|e1hifz^ z%;oks-zXjQ!xAs)(APV?^dzMHV0Jt4fS{`P+!qBBHgCp$@p}TX`iMQOScKIk*_x*= zQFGk`pSM~|ANy76NEM0QCG@J_bOERazqP-|^V&GlIhy;;VX0IT(&KF|=u@U>00#s8 zDT>@z%-KufMURXBr4>L*?< zN=KNZk@?e2uWxuhYvLwAXDI$r0kJ0HAqILivYNhMk^ z7-#uI)#|qrnA7c60et-+yBxv-&tt<6C`1+A3K~H7%w9jSlF^xR`Y@?e+$0!!AU^Xf zft7dO*qm=6dpLfv=D6tzaWz4GYYMyYo(CMbJa`<#?=g%~`yfFU-ibHv*{|0OIu=Y7yEhOI4Nx>sk7=P3;HT}3@j@AK z25*6C18syx zT$YbBWvve=TYdLl0CR}Yy|xh8w};O$$I%5*$H+#4ZdnBWToK%Fi0&>Mi}jJ|G2#}U zRrJDaM$PE|Xrm9*9dG8an2^$bb^x!$ZfpJu4C`fq9 zv*~$1J4YKd19=JOSRh!u`N|gHX`uws*&cRiX+PEs`v>4 z+K7X2>`QC-ojnkvspL}~`gsNOUXB1%zqEcx#n0hU+q&|bJcR>ZNBW&j@0eQW+nig} zkwxnbk3@w5?J7Qu>IfN4R<*F1o?fss8b)r3`7%ELh#p@)HT%~^q%Cc)@4&QyoxMa) zK+c2n1AS#x%eR1P#NYMqe(Cj?;2KzU<2PEIU{A=?ZAB{J$%S0Xfu+{veFrC4R472Y zaeKit>wXtpH3t?m|M#OgQ@|DD+i3kgoHEPwWRj?;MLo8Z?VzrA6ltOjE&Sr& z+6K(RW2_|PGfPrWlNVc0^22kK4&NqJ4<-=+a;pG|;e zkD#@w)p3!|JRlusaq$H&N~j~z)Av7(49;|z;)3;R@GikQr1T_pLF#y8c@~Fr3jP&S{&~K=sZ6*|8jHzK>pJ*4Wx}F~tyB6PVaSTmU}&|N-{uXwos=D| z+jut)F|I>nfY*(CjK9v?SAKM6t=mPvm`zlA4L>W+sob9o1?Z3o`x14b@nO1Tr|)>i z-IRI~Jy7%PY>waMxt^cOx*>0aMoXkVw#l+!BHQCJxi1@26#U@6#NN8B+>2-d!N{w2 zJ@#ruo8Els#2=3!r5rTGgAjik2oI2UTF8$Imf~h#sIE3OXw}N@Rtcs3Q@E>Hz@%6* zr30`<_=Q2u@ zf=TaxUUD2vEs)KsgFjYq6=1MnGPD zubzT0K;x$Vs@Bp6V>&(&fi!wo-{+AFaqW=_%Sl=WlAijFSJAkZ0^`EDQu4TGg3g5QFV%WS&J#x!7*W*xeNL3u&K{4%qeN~d>_#!FAOL=rz1W`bF{=Z z@c3}t_%E470!)!->%N#nLBm&p?Jtr4TK)hgQP-D9YCTbBi<1bO*tHr$dRQ8Y|H5mOd@fa5qa+Fw1l+|pC>nG;_yPUws7 z?|meY_p23By+pr`@S>m>!Jz#uVm!aT=$az-{+R?%)~7Jxg2%;iLym!6Hm^he_?+;?^avWR?{Dkmf_5};nX=KCua_l*7VC^nK?t2hE|hL#${WG~VVMG79ykyij+{ESL7#%#hC& zg>pyLjCI|{S>=%-+ER+~boWO<+2nTr@n8C>bo;prUyC>^Sd!#Q9?LSmkgWD1eh$QM zst3kpn;B_C`Ejqt7hH=#Z7yV0(4)ti-S1=9z#4->7MsmrRHvcl75=IaytD2RUk4BaJzai z@xcxJf}0!^sW{W3em(pZhKU3PI(HoJ;9EWqm|Wfby=|5qdL{Kw2TFRce1EO22eZHa zZ$y>WWEo%(MNhOT;K&G5p1*NF9v55tDJ`5`0RQ=fJorgA+G=uXG z3H-;gQyhSOup%>9w6mgKgD-(Nyvhy{XG6kD|KOp;GJZcij)d=2ftvOW6y{1qqEE+z zZ6Oi1HO^AGK8`nTR9msC)z^vd)BWn7>uZpA><=v|397Y_A;>{)u?<(hujy{)dBCE^ zVKD!@VXPV|OI6rH#c@5F4(_c-lM(UO#dVpf!)QGHJ)4H_G3Vhp-cnPPv9vz1myF#9Gh6MtU_?th0X{UEa;gO&-Nkq#m2V_b3E??s zqE)UKtP`W%bvLH?M)=9@&H0|?^Jisex#QM&kI!fA4Ch`oj8^d4Lv8F{5%!!I&Yn>s zll$-C@PYiJWJJ6AR$|8?lS9=$z+%E2MFBwzhLmhha`2JV& zr?OKJ!>ObG0{})QadMo^22mRO{!oj=n08eMb&IIHMyo=aSSSm8e;dTPykIu(=Q-$P46uCvWL zF@CJB8O1unLye$xgx5$AAHN`Kzs%3`yL&fxssXI}U^`upV_yea6#1!sd_w5t!oXsq zt7!}qWOqwALc59vKD;qhzLO?Y*jK9HE8o@^JNcUWxn~uMw<}bnd);@!8QhXCGQ=&s zZ?~gt3MKo5v=*VVvShNzi+Dn`Z`~~UxlOy)Ag4bO0=~NCFW821P9XvxzYoP{1*xid zBgbCLaTi`eAjJEZogDm=y`E3+H&W++f!heS9c=u3t`R8M4X>|yF*5v+B;5)+EK?=A zZn^Ok9m_*Q;d&89e9!4)ZV33(Q^ zon^#r9}G+znq)8KQSP6=+v!m`!YR=0Q{p1PbqrS5Bwp4#c{&3>a~wEX3Sg-*E)!en zkL38Ww^3{DgN8%Bk z7iehm=tO%77D;SVadRR?0_yJ`M^PS~;>*7O7$P{soOrz=u0+cfc4L8R_4ng$x)%wk ziGK=DAWRCAea<|#8CF}Hw6w7?)|a|0i z%PYhFEP3&}SgBqt!Pau!ai&W~17u^mO*4F-9<;^; zIbz+X4E1#{FKb)T%!kJ`A;>d4JuATUmn53s+NjMKf-B6D2Ou8w-x-N2>KvOUeyS zpJaY19yTXQm_rs@fuHG@x3Ksi`1jS2o2Fc+LQON>=f;{~xdd^z2rf0fs1K|&QSA9jn z$Jym4sk(`~_xu1pVc`lwHNZFr7$}_fkQzP+oW!`7?{Df{zWgb6iwM`&QDW|tID8*H zP}ObLAdmp=UI>fsOZ;}h6{-+zh$tV=fz%W2#m8vQKv%HpMEmY~dsq5qv+h#?#05oN z$(%->N_4)*{)I!&(LG_-o?XhjVQ{E)n?ZSA6O>L~N0Fh5$7CONkbe$wPRKKC#Ul z&Y&J)7*o)NK)>sLxLn_7xESz+Rc|ASHCcrZ)$@uve4y~hTVj>UMYks`f7D1lk7r^A zeWr)C-?Da~r&L#&Bmucx`jre^h9DhsLXqHB88KggIfKsfrwXmyc72r@SgnP zd+eGAICHB7ZCnGZ2>keaNpfhfK}R_`di5>gl>Sa1gj1$j7@Psy_`m%jYJw)k9AIku z6^L4H^0@C1g~rwi1|9o5FwNY3A5$sx3O=W3zd5Vp+t7oi8XsYueZ+-<`>F^E`xu|8 z+tCcJKGZIiNr;=QnY!MFqZ{YV`34_ImOMUb*PoVRj3)gKjAub(4zPXtq=-C~iN;xv zceY9Sikyx6QmDpMn}DEM+g9D5>umfSjg-n5`sBN zu7Q9u8Z%*dLE_Po(LF^H9eVVr_V_soTvb*V`PW`4>#5S`VoZ<!lXJd!x+mZ` zXmYu??I^m~^_u?S!r-H)Fg^zN9|`ZeoPlEl&aeQlCSZ~(Vy4YaWXAB?gI#e*7@Xb% zEU)_vn>kPa$HFN?EVoUzo4%@No(zG3it~>H%S|F~L67>wT!l+W2K(@Jr4Y_eeZT#@ zyr&-nduakyR$leB2i+yGPT?KWv^O3JnhyXf2B49OGq#ZtzlZEgeq*^d*#fbWfRXXl z?#Qk`$L|k{h{@xCXa=lrvdDnSzhBAr%lJsudhA>d~q7 z#rwC1VD=mL`>JN3KXU0-Yaf`p2Z4e!oXAfsJTolf*R+2JAxbzgh8g{ex3CX#U!)Ui zxS8w$wii%7_bc}fe_znh_~b|xt_XRaQfB^6ovHB$dKCo_fuH+M>2jOPBx{Fjy0Q=< zwjK7a9Mr>rmoE^>o?=;}9~PM{QVTfs9nTE*LGg8h{###QURc-}m=0uRP5Ukmf1+b~ z(A5Eltm?)Bk#cv)F!D$&4EZ6hoFMU|eZ60Wj7fR%4S6TG1w&H4gW>#q9HwB1#@^RU zHzrj!x1FnAFqawx@;Hd3%F&0%`SS$q`3U6+w1WyF=DI%bHaUPBxi;6&?@YKc)S32V)HF3v74b}+ zAwwDPRK*MJs1}b!Gu$aj5CmTI`@YljheY!;9C#-R_zokP!l1N<;=tILw@mn2iYC+))nCmmS@Z}E_VcD0z<{^jT7#BoFv7u1J z`;E1?y;&d&^AapQOnfAbWeZ-keo`s&LJ+TJ*VAGBoJzgoB`u?wv-EyE-}!gRf=+~x zkPj$Tj>`E9;17YOOx?Eno&l!Nx-uKVy)9@Y~-u%>wo6aX>so#R@5 zt7m!e?T@yPPB0G`-S4$AH|e`A*BWXZIi~TX(zq}3{undDG~Oul2h%~z;tF`cJR`QP z@MWhnYugOU?6+@u3IbR?KggbikYgGs6c&%ja@zu8ba+a6X!T2bXqaX6rjog1P+spp zw2dI?WHl_^Kj77)RXBkS0~bNq6epG+#)vr-bU`NXg-S4!ZwI_%+sp=zVw~DzEa)^H z4tu73_3kd{v2l92UnIl5ij&rrxiwiRUCZB|G(I%+p@%%{h}S_wv`}p=VGn9u3xFNc zB+$jlh)CTJNntO73=~ZQe?_(KIMDazn(!mF!lX}+O8^sodgxB;3pQ3?L*ExwM$ny= z>PdlQUn12@dS16AJ7~|wPW=7K;5I#Z3iOE8lo^}wwiL1U-gkT7c){`rWPMR{N~bbZ zKvcr?XXz(yv;m9Q$BiV`;%aHBN`)1nidD7`sy?|sh+tX(PbuR&1jz`^f*b((=KJc- zf-p`bDu`3m4B@#!2py+*wGiHjLnr|f^u&Yc?x9+2*2$ZrY3!p_QocGnv2WeEV81grbBE+-fP@pGU z&F+Adj1v@LC0F_7Ll4HV^N8SZTfg`Tt9%fV$={k(mRq-^sTB!omA8<%o}}}fPLY^D zgAfFOk=2x6iA+{VV!5KdB5FYQFQ?0sFgt(64q7_wE4opHd`H$|6Cp{+X7OiKdYOiL z-y7gZMChP0ZnO^Yq2jSt?mODtXFm+n=X1V;juCK{9NhfZ46F`9odV#)^%XMdj%4Ig zi9Fs&(D3TV=Z`od<9td=jLC6#f~n;*^{I1T!#S_ts}>^Gqp(k!`&qYh;q?Lk0ks@P z3N)l3>y}hRszE4Mg_vzG2$rgUI|RAe=6*W%t$YBm$$n#L&*HRSk-E*xVRtBe(c0L~ z1bKa5kh%odM~B3MY7oTwEkSAS30XR>ZCuYeb-D9-I(XF?8CUzW zZbW`~2FVW1sU_rx1@vC4Pd;b3EaZ2i>>@Ka1z|2`1HK{bb+tQNDzs`^dRZYEV8jH{ z5aL(~4eL0w*Jmtpt>m3G=sT1jwxL%Tp0xDN)%Ht|4Dsdf>!K@mN^NU+zpe3DO1rJY zg}KWW+Kt>=rp*BD!oD@C#VG>I`h%;HdzBOX)75@?xWL*6!{A5rnm;)5CfP{||Hf-$ z5vfsXUUmnDF&5XNg9qG2MfCLQe>WB?{A*&X`%S$-y$0F-*CAf?4ZQ-)I+D_$5sW<} z9(u$4jw@-eO`W0R&!KfT3chMyAA~OSiU)ilaR4`p4Y~SGBT(uNhxOqw43!LN+{VF3 z;qR9o$t$RQS^`4_8<@a&5!mW+iC{M(qu7lE*k<6{TUT0NT-kriG@z1(Ky0EvZX>J0 ztXYD~i`}6BED*onu`hCU754FwLlpqIcZS*c*@DH1Xp8XD7aW{-XrA5bzq@Q3R0Z4zuV%p;`SR(#Xy)5+>ky@^0%XI> zX-SAZ;#k6pB5&PwF>1xvTNi0EHW>6ayQ3XW=SOs5v~SuXNv^Di<^W7A9p=SnAL5kY zI$f45v41O~WL`SEjL0^>-e<@va$6Hw4Hpd#lj-KnH<;|3xdN5DO2T*fA&<|d;ms(z zl6=(lP<~SbR=UR^DF#@D(R+wr`!qb#q$lCwnuk%v>{+X0aB@!sIbGW};E<@XvGns& zWx$(pU*g7PyofdxT5Pw5mDUGYCa>FK{}G8>GD{FZom}hO_Ko*Lk4xRa^*LVfLi=3$ zDLxr>6Q~y_eq~?ednDV)<5<1^Vw*qK*sg3n0&+;Id!^PJnEyg5>YQNOxb(pW#RDZ^flS)OOo0v?hYfT1WRl}| zRhFM(IA4O{#Rkdlr5Ni`8kH<*$@KiKrA>sk0dfHKhb8w#*vQ+NT!QOxFjhRNZ~uxF zGQbHk0eaQQ#5=#)er1inGe=&EqVlj6Tl#0_Mt36e-wvup^dkbTs4w{CaSq`*Dk+Bi z&bzW8lfTBLc!*11IwDFnC&XGEuXcNSoT(NTg9+(w;AKIfL_|ICsPUM`F!e&W6O_y$ z(S(6Bw8w@%37%{M&{aVUm4Qsg7oC3r!|}fva8Klz-o=6S=03;a8@99+bnAPIKz=|T z1Rz&{$^lq#K{l0l+$TSVlJ5+lETg=yHM&2nkeL@Rv%6UJc^^LEs5(85a8S-MhG8U5 zK(waeBE`dD3|2R?;&bGqXMg_L<3i|d=?_oy5?TCxZ(;y@@{`QII-b!dzPwJ2*Uk5) zY-_3A#FF-5WV0Yp-1~dg)bcqQdXI~@54@NV_fg<8cR(WO2n;a9a6X}f_^R});F9}5 zxPk@p7F(DFE-+<@HLAgejG#~vI;caEV@CX8XP|L|(*(cclQgejgi}Z~+NTnzi68Cc z0{f@1zxJI>H~BQwMWTK0Vdu3V``pdPxuNqN_* z1_5NObvqnvJTEGj@TuOdTSAE~`3(-r_f?OFv_R~`KvCsgv6p>LJe{Y@^Qzb*D zgB2a|dLJF229IP>f&%U$1ecA1!UI8$tAGqq0E@*>^VbrO4^TS92^@K?d%iDS>O+-* z?qw@SD;CAL7$l$j#Wi~1sTX&mppq$XNV|Bn_Za2| zN(6QR*=0R?v3n5!mk&uW*}iO4Y4^9XSTmzXd>T!e=2GL?>%48 zD;ilRveShm9O|$_F#Q-JwJ7b(C=K1Yo#aStYg zIS{{xsp>RcP|DH)Q-OT~wR$*Fa$S1j8eF;%6#!)Yp-#P@?WfALY@P-HI(&T{b%xUn zD58hX{X!x1VH$}S_5?Qh^9Q>3>pLkd6|uYq|bl#(eLHcf%znB`Oxy@k?m0l{p`bai{sg8x5tnWll)i~ zk$SHW!-Tcgl_KXXawaA+0 zOgv=0^%iHca0==bjb9KNzAB6mgDT~lDIJU%5m7v4d~Qvl-wZ!dSq{BxXxzGA5&u1_*f8d4ay33Vl;+ns@gD>+JaZeE1 z1q6Pv6bD&Udzg(iIR$$SOYcB@onkd_&r_zW*tgqBZB*G+e*>iHd80Z_?b1Vjb`8GA z%6G>yL{ZJipNQkxR|uZ?VK+{Z5(EK@YIxj^MZQSF5i1H9?oVP3oIJudY9^q5rFeip zO>!z1MW5e1{EV|5<@qGIk|G34AJ~eZX}@M485nkE$^I4+teR7YyxF%aThUjDq9Yt; z1^n$zs#*KIzm~y?P?g#CY1B2Uqop@QSmgl-t3i3G)fSPNLxg2={6X`cJ4`K7$smOA zJO>&djwGbJaQO*E%~5>%Fr*r6B~m?)QA)7p@+QKGEwZm-v}c!QyjnO~t53gd((}vXeIVooY^};k_)r7z|Jp@_ zlYO)8U|GNg${r%-UnEMsWn-JY1^Lk7a*m`jaSgif@-^RBfa*Jfu^(W1yp#heQo(q$ zI33XP-Y&IZu#1m?kk+B1T&Yy~&o5{!XzEx6joo|BZmSGY`A;-{NQ(lO0f#Lif%02F zo_RF=3Wi|>MX_4m{Snlj`&!3*uY-FVIzxjO0o{vG31FygCYDA8`-q;bjHv_LtxEls z;j-^-$m=($5&wk!Cx-5L|JEKE=V_p{?|X_f|M^bBVQKx`eM$q$1C>&USE{AFF+xxd zU-0g}l|V!ses#P9&}X;P(~eh(x%yvdf<1sBp&!d3kueeIBjgI&IhW9tD`y&oko3VG zexVn!U005|{^C_YMMoNJs4Da@P4ZMdpCESL$BH|Ca6Md?7quDOwQi2mz5rf$8p{xi zA@cszAVUMF5w4ICHy4gwsbzKTzNv$z3RFSu?*sU88v+*3vg}Il z>LW@G$VF&#UHx@$Qy7|sFy}KWi}2S{CjTeY)$UXXWquMW&frl@Lng=A0 z)ms%B`RC|3w32YTwUeNa1{-RJC8+id{vy14tTr~cukzE+pl`+FW%w_|P^T-PA(yGE zHdrArBp`o-_8T^cFOCf=ggn%M`hcjA14lIOqH7L_j`4N~P|^h;dVyQ&aR)AKsUC-2 zKl#;r)J5Q=Ma6~7(YmCJL8_9f%^d1pLN{Y&&v=`TSR@$>%JK~=Ywi7ay@SQY-o)z4 zjOIv<;TsX`G{9w#eU8&O9nWm>*Sp5vt;qU5>%)3K?B}=*W?a{pJ@lmF{U?%duJD2 zJD|1%&3Tt-qeX=m+bUxg_qE0%)XI0+`9wAwvj`rgAap%vNpg_z(=Md(7R*#GUY1+R z)k5DfMuUL4CZUGCPXQJ-h6$SsqA;I39O*Uk&se7VaZo-pk&ymyes2=t z)*ukry>Zv;5etjld}!u7ZzKo5k_5>Y)}5m-cWp>7j&mxRpSa(u$m_l}50@|n4V(Pp zYph;T9;S;i6T*DMII}PJypVK#p$f|JhN_DI#Pe;QE=>B1Rc~@|8T`IYa1$SjK3{oK z$cuSF%Ja>1k1_FaFONC1o;*Ds%=NPm_=2!q|4 zZ%{sRfW3senHoG~PJlx+?4=g;xh4OGvbWM*#Z<}xj`6Aj#R?FREV$!OjkFRZJP&f~ z%yKS#orRX1=t>DY8{^N9*eK=n6+%5vSpP43v}xEBG|k~;Tkj0%xA*R8!OGht4~wtV zk_Udy6@vWWeNm8z=lM59ufm~3+Z?p;maXBkImphfrcfyFQw5 z^Nn5EFhzRXXLzv+!e-~9v{wGZ(qX4+X*t5~lUdbUD7!QuUEBS3`Aj{|)?7|^H5D%Q zPx90R>7ij}LXwKR!yeUJHD+MBMbO4=XF9O8po4~_LC)+ImhR?)a;xdH2zXSnC>eZB zkcR=GH}B$|+)t2M^I`z0#TN=7>jc@nAP5Lixa@st$kzx(yX@dRiNgKBsUH(bq|YS- zF)lbaP;gM!h>1l`C*|Xe@{A0r=Qizn-~svuaRZg#);zf zAdkiUe1~R8bRmPv2)2LvM|=WnAEg%oxG8KTF37%H4+S!Lp|1(SG0$!k%V0T$90{92 zQ1FBDtsk`)=rWBr?*8zm;H}%Qx5V%s^@b_e*`T zgo3NxfnP3&^XTRc-!F|~NOgcVs!1a*zbQzOaO%rU zKKga(PouaC|4Mf`57HXbWz>`FB2%MLj(EaTk|F#wy-XgusGE7@pczT3qCFmM4MlKb z3Qa9Y_uBG+>a1V{XY`@D2Tl%$C^^4w7LdF@4+xov;T`|h@gaNW6>ST?#2yzi3JFjx%V)3ayv}91)vo8%jvJYx23{DeD?S?f?;a@; z-U%I<=E}_bi@r(X`XcV(zzW)aNm3xHKiTK8e$mg*jR`|1FLJk5_k)18i|6AjvV`OO zb$fK50svb4l8Pj#uM-P>6PO#F4N zve<-J4(0XCk@kz@&tX1(LM$^WB3>;FADjg0Nb8yhTLrA5m(#m(dmV1CdgDP!$1WxX z5Jt%m7}Mv{9(PH-lO`*sRg%xE@preI+tu64cqm^4jR|XL!#2};Fm52_Pa*-~gK=)Q z_rA&Ie_LC(!XO7!Q=xaN|3>dG)pZg4pvTeMFhJi*Z3xN*BwPoYDwwE&U$uXr?E&`f zIV)^|!((9A%*;0t837qOOHG+V$0i$fFhRM%jQR#bQRq`c->H)kBQqgW8v+OeBdG2qi*Bv>`L&zl>ynK|1-vA+N%)18xr(SzZEs|M` z!_Jeo(n|A_nl4@+GH(vk`3h~&=Pl-TVpT>6DlIU5F52YDH?S@P#r<8Ew*tn_s-m+Mi>#g84{UrtVcT;w(rPr`oq}J!F6cqrJ$iOwwB%V~_{rpO@rb=5 zpfK-s`c~%;HSBmBj-HptO|ZxAOM+yHIT;KW^%TG`&-c~oh8e}U0 zaEcysy*`cEI&|o#(Ot#^k(iJ0DF*WmKyV8A>Lmc3sGH#8<2e0gL~Qj7^d1uqRIiot8cB7#oQ^#*R`?_6+*t3G$ zYq+C8hvTrh5_ApE*9ZfT@Or1SrF$9$RGx_BZcjmDn;zFJxc&3NSHvE}AQW#nw`FR4`W;DYr zXDd{@P|}~o+uaUP-7Nm3vE^wm#4ENc`>*^xF<28Xk&Q_TMBH~oUk^3cJq~tI)$sQs zHx!@Y$Ye+=e}bQumXhkrtWIIS`U^7}m{Y$eJDXR!I9)^EBU;qrZ!NCF5vf;g9q54Jgu$MpFE;;b2gY8YJrG_~G(`zErDqgh%y(cTw;{9GpDEs3+4V1)&}+ zy4&@Cp6*(pHPv~~M6BB2AcpU0d%zs-zk*-&VD=jmZw@xxh@F3!%DIJxrd||7R;Dk} z{v(DpE9?i2<)Aw+8?#D;FbCk6qU&ZL=@$!?v#s?h;0wv?Io$ed487D#(jrgg)g)6( z#6u@|V(B8BuR6h>P3V*YUdJwY$;;VzkhE1k2V}w&_M8k(OyKJH7;qPmz71~JUi6ok~T)@|la`Dw$KP%pKrl;1xO2|V6 zP*8&@w8^nm{0B0tuFrzJ-!Ow+_2a@dZO;)|nr&ChzVh0%?TcUpYSpeJbcv8FLq2VQCn+(8a zuAr&hf5_o`)itux7^KZNMFRoMa@4dtSQBy6HBEU$b@C5gjcUikvN)go1*>S0NHWy& zuX_&2@3^Le9(2O~%L z?|g%vl$fM6gvV*aDd+CG->P;Xgr9c%sgKq#3{*}4eSkI+fQz#C-4L8FxR6&_fFs$7cr0$zDO$a+&70aE5B zowOIY58&mNk3zF4%hN_Oo%Y*DUczy<%kgUww^MVQvz9=-d`{W{BCGO!9({7qDVq=W zBhM;It)7Ik0DkRT@Q`)Jat0us`-a%vJMH0qxrdfQ4_~3GLtit0KbCbB#Qb57cJnr1 ztoD%NB-an;0u<+T-di5McDm}|D^x8fhqzO}BS~G{%7WOmsY(dcw1IkdpPD$sWIHQL z<|&4XC~CfE2Lm|F=LzY|Kq)P|i`q;T&#Vfy2b z^g9+CLb?1X_eMSf23u0hM(|&GRmS`!(~wp_N0bw_1+Q34A=)RP#V(31lha9m44J@Xwys`AAK_1NlYE9oVN8 z>8K_5LC(*+e(9k^WMtNXTH(Bu)6c#6m{X`TqcS|TG$XJNUDQ!naIXlGTzR`6cepzO zfA>wwE)QJ<{>!IBoaJVhFobQry{V~L~R8B zkC&!tbWGLncPsgB1;1x5No_yPGOWT5(Zh@h>$r2TkqEe6%CVe6&QWtr>U(eMOWR|y z^F*aGE{+QLy^PbKG!D&A)ayv@C;&f`>+E|9U-k3iJx}=#`9j9g_fpXh9sgYZlB5>GsoGX(U&U+4RdP_~Vy z&?NgEiws9kdTOE};X`s&@dP8WJ}CPFTYSpQE3)?~YTgSX=L#%e;ewnIlu%uAf2G{W z#X|p`7{nDLd9h&$L$p`G88v|j_*)Fp6H+v6E&HTICgxKf-Cy@t(wwq{qH!~7@asyK z=UO8`$vt*y*;6!CiaJYe((8WIzGu7${_OKS&*J<^Q_rGxTY%ape1q(wi`fxq4J7Ek zUTJX^P^yp0EL4Mhfz&(PVLDC-;Y&V_IxT#OYBSP1m+?R=H$UdJbYJ|{E^l1Q!p|*D(>xT|Yn|KZ z;Ig@1b?U#)D-!_y&)A7HgJO~UQw+-I`VTvR!cPAS5 zN?1q>3s*SxzwUd*c8tawjJ&d6e!7~vNVhcm=Y@dM|74>U2?V7MG^hO3(Y;}@GoM)R z3{Mw=--dKmD3Z!i;Oa6jLXU&@A!l6s`QmLXFWl8yop^wDGA?onVTZo4}+O=*kt`G<^*ruE^$F|ciN zv&&J7Pc9mJw)hcU8VNL0QgC`u7KokjGHKq1y>y9z+?Ks*6fy97t;ZhI0e}?XZ@a_b z-{cmAZaFlr&vp50hwgK_sL|$r44Qrazn^Yx`R-jWtf7p$J`|SDK!uN5B`<=Enq|C!|7WkUj>&vMt%u&9N)OAXIMO0T7!@l?^6QP z_B@4GYJblUQHBl&|42#@;VT`u64l#;z`W$|`bOB6?ZvXuw09m^szQRe7o@pf{=GXvngDg$DUxB%j=2x2_5B6anPH7V$)fU)&R{(qX!o`Cq)Glu+u-7wSS_VZ zpQYJ&eUIL@q8G7nU{MMBp^~@>dm%m$BZS4cUHa`l&cbQd<(!$c^UC7+h*d!%W8iv*L2;NWWwc(1&{diDGxb{&rVUchv}tRDywF=ET82yfd>xOF z|BeAV5c)D{0e=q&T#aK9-M;*s`@UPE{Vlt^Pj+`#4Kfk5pcA}R*gh)k$SJ_(S0mCc z*bVMF1Srz?V@BOyU?dw8z2L3rw1G2nlk)Z#HNGDJBI)ti5*TnOuSajoEC%9``#H29 zK*G~5Yc(1&y_@*SO|!Ru_ux`c|c?aGPi;9f->95z2;wKUY_Qdi|z$a$3;j?a!iMK*tau9OrasKo>() z!iWc6ay46*bNky194EZMQ=BYsvORMtGqFb~IBc&o=uL;2CbTKH#DjcIK^^r(xI0rtVIfnebi+WNsKp7)P~tOmTiIiQ?}YBO27+SQp z!<}^!Ji@ST0NiEoP2N3OTeE>Irb_(I?AvV5i11kdtR2xCM?(*z0oVU>KOs#675Qpw ze%f7l2T{@-z3l35Z$SIzAua+5@p?%7h9pBt&u4~2UvqgQJ@%Fhx9e5d%ZfeYQE$e& zyjG+i3R*GLmy6D{QbVL?3`h{N2#oP$!e1;B9&=u!P*0IIb4oNPK@B{+PS^LgUlpN> zrFZTmIq124CCkpBhXbIM%;Um%n@J0$rUw@yQGQ)<0%-Wc)ju!Nv{U9Td8c22mP9bJ z?y`tSaa114cHB=vUly$9!QHx#J@~_oAKjs}vz8s;rG;_-e2Vsav}TVLltD&#^j)Ax zdk#A3eVr2fnr|K*=)**=*eLHZ6*8a&|a4QHKunzQEal-%->*jC(I&wt1( z=cR#~7x)KWvC7;jy#u!;hrguG5Wj{!S#pcOiES1?H8k0U>UlEhY*ao?Pe@<-ldM3h3EK@K;M@d(9ium&+EYz`h5ZCUhh;0Grl#Rd;{^Up zD!zXrR&WAjNxxw)&jmIGMl*t4e22t;-vYHJhoAfY zUWNU;Ql6ZJ0XeWNPYFfDe(;0(@j7z+N65?N2A#rivLYf9`Q5s+UkmME=BluVD<5TfDCfIN}t2_LoO?{WAa7z+nELgRwwIuRvT)X2Iy28SbPq*y!)8$d0QY?o zKk`H;!*57`Z=JUH>$3cALs&|Hi6)Q(c;3*aWL)?rq69nt*t~N2^QJD(zu;seO&72E z#OLetas_h~MgwEEzYGMLga^_SNRW7&PsMZ`@d%Cc-F!^XJ$x4IB=IBBLtQMj&rETU zaNfa_*=@gLzInSSf25lbbUO;!jajtA=kY8@dOhKUmQ8&4;*qa26eov^fWqszATNlR ztbTdCPz1XGR)kG3_9TrN>R0o)ER*+!ZO5_zw4sRsQBGk<4j|kJk_m4rnalKXUB^f> zPx2rM?v7Blta;t)PIM1dhme-9EvKAR4( z?n(Zf+~Z~WT&ub+45iAKeZEME;*8@8lGRp2tPHE5IZy*Sv}D|J%=galerQAhNc<)G z%%!r4C`+~P%M>zT8@u;O)hT+7bShji;{?95*6;7|1|z9c4`GyPQZQBDN8qgt%9XYb z%rctGU~&$-NDtk!;Op*p6BKwP^k&g>W%9j5C+5b>M(;rPn?&?1IfA=2(J*{~-SmPE z)Zt?zMtn?!$6j-EZK|3=w;w5dCX*p5pxQD)z+{fkn9$$`tk;g*15cau+fYNQocMJK z=utTu@lrGD@)6Mue%(4v^uAZQYSN&YjH_G7qp+97%tz(3?m|YJY8XE{z_o?8zJuKs z6R$5CnESsmj8PaH{yn1j{WZBYgfE9LFptcU2W$_BE8b-!S5g5zZ$<6a8231 z@M$hAWBAardSVK2Z?1(R1Eu%oSpg39$b zqCxREvm||%=FZj^be8Swi*@FTD*s z<+hBKsPsKTwN6Pdy(s7#+>Pu0`z3%)c#u>f%6*8y=Dby)$3oug%Xg|=zlhK2Tl(%{ zVY7(Sj6j&*z1&Vh^mUon))rOQAJb-^zcj;uPp5KPb3ku&Uuchs7ow}OABjwVA+NK$ zR-GhbGA4@I8GYo{%8Ppa)RN8n0j#B=8=x@X{rd~#y^KedOyEu z$Sun=sMGnFEh9c7;|3fmXYJj9s*jnP+hO&yf(gL6>)%a)CPLey zJ)qVb44-}K0@5%vzstE~Y*B37kOiMLeRm*Q(V=ka<}IAs4xF!RCc&aRMl!?wh9_oa zj@_qba|4IA&8iGe(3O=ksFLrvgI&jpzkPKN4R77x=(U;Y%^^<=pi|)$5lXJxcS&qXvaIACmy)&^Qyoc2u* zb>eg7vfNl#7B|w}oJC31o^6kup2fe3Cs;?x+2k=^g>$U2JiKhJ0praNF&ob|Ibl=s zZdpGEJ(oHasF|6KIONdGfL9NZDdU011f3oi>U!A$8ub`|auAA5eU6Neo8F{3kcUgx z8Nt4ISV|H7Wdpj_Mt#G^f?jYap9{JU0QADBAw2>JkTc1_rUSxuT}4*`b#PVz`|r1; ze(={eJRy(O#E?wf&mVDmY*^V$3-~pEdy^v)>XF^ljj<~>%;b9gO-oVv4WV)6YfZYp zZD6HpaJ?yVKhw#+9j&J*0{H?llfj&D+B>Ez92P5(TSKmH?3agAHmEegDf(Ta*#Q5d zz?pDWQM}h%3HkBhOY8@~;ZU#XAI{{F0ntOz@MC`PMFwo?5$u>6*QiWr*E^q)a;kSS zs8!Gz4kH;`90n`iT|pQsse!_u(TI^<;qag#gnh>@#c zv5(XEKwd)zfxp{<+nhHFLQ=0>JM8W$z8H@0(Z$o@&VK8YVVHVCzSa%lX6>-=>Sk0t zhlkmB;Q9cffPKy^JeDu2$!6f(9rpH!uAq#73TgTE`KqRQFa^*blI8IcEo?sjtm~6p z4pW57+ao@mq;cm}@)4=I+<26}CJF{xVxgi>MT#BF?vMtIxnShUDXw)C-3a|${k6Tw zFN9pKw|vyrVnHyM=PSg|SkMJ6N?PgONWtL6VQ)g}!vm(FJpP>xwEm;orvzwsru1H4%{MaqIc^x!9m4WU3i9l8cxu)r=^-V`i*|Yd0 z3+hE$AY$E0EZfT`Bfmet%*ePp#~T{j0n)czNyV@zvpl>hymGjg7bYs@`_s)VALvdV zqzEW0Yk4hSCeofSw{v)tT3$Q(p5%hOY}X^kedjpgxKP6z5t!hKJJZ`8GPN05RnwrY zx(7(%yVt=NK@v{hgh_?4kh1&g#Sk10$feX@#37}mj1yaW63q8Rp@J(MKkx94RXd)# zPF`S=c~!RRtPZ`)(}V=}+NPJgLfVR}Wf1GS><7Ab);W1=_WlT91rb7JV;N9rc4)fU z6>c2d=~6{}uaG4xz&_%CV|CvFqu}mg7PvtKHmNxczVrL|Wp37Zma+GdhOD~=+Ov&9xU^lQ7V$CvO);Jt@O&WJ2{X$6{VCE5^#?KCq{Z|v>TKLk` zdl3&^x0lBkHVh;ys2y25O1GnC*{{R8_kfKRG@P(R$9PM`Q#ccsi7QH}8=mv?CC6bK z4>KK+2Qby8v~n>WBo&o?kBUtV$o$pwuh1WIa6w-y1m7SbR|E8Bf;86dR%Im9x0$E9~^{JA#BX0 zGW8cAQb68?4~)Z{1|6pp+*d>S>TfurAM@$Y{=dafA{!zR-Yh6~K{@7k9*=~5ri5Ty z9{DbweL);aPkf&6kV&K>ogcD!a7F|W1>lrksw2X=s#9>3>f_A)+*d8^otK3zwyH-T zA1@xgFL(GT{Pklj$UE|eTQO+o-y0ZNNR_dM!PXkdmwC|*_Y?3g(<+2%IUy}2YupDD z+R-p>AU4kP?ECWbr`hJMK*7XJ{n2?S%Fvk*gg^&-IFHRdX3FzL1H)Cpz5N{3G8hgV z08^L<-8>yn)+IaSih-jEIJl_wVzVq;lx?SU%nl^R4xHHE=fHn9!W;a(WV3MF7hKTQ zux0K9AnlKW_RTt{C!bT~_JW?myM*YyFW?0%4%fd>XQZNv)Zo9p(#y#c==O)J_qM2O-hALzXkE1;1|{!rO+zGT z%aR9jxl$vfBnb54e^SBI!1Jsu@)uytvq9*#$XrEr-x~;|Q%AGXb1_ELjK}_)XzJ`z zVH(Gm?nB)XFqu0pAG|*pl7Uzl`n`3;b=H0_6VrPiEYj4Nnd7Ip9ZoZiZ>v6hVBQe8 zPBgVzpv%}H-mumKl~C9nkJ4qFu^rRM#&UGTI)y6K7!4S<>%=_qFlnJuJ4dGy4`o<) zyXs4-Sh?cy8EttLqWhXF+NQ^T&m!d>^(3cUC;fsDW^JBI7iYh%9>tsJRezPl5OixB zt2jb(FBkj*;&c!TIb$$+b6bpP`F|-5D>Bgbk0u06@G#%!DPYWTd>kaIq12z`FmLhj zbl>F}PtN`=Z#t7>_Qf^BDlvDhe9LqLIeLZH{Snt-kv;g0XAMOA3$SU2(@*8sBP3P5 z5Ti8r=f0>*1nowiWPL0dV%2coQ;l@arzs}4lcGj24g>fb}_Zr=lRcmAM%Gq z+&m26ZQJ;w2{djNj<9KxGXts5wbRaX^hOHM<&HI>?oZPv?w|{VeXz&K1R?`5pcHvi z>g)Q()2%T;(VrcgRc+!#>vVXGd8|vN;Hbg~{bwn>hEL&Ilh>a-J^ms~Q5r7yLCDf& z>KtFTA9NM_<|RhSQ)Bw%NM zWM`E066F7)UlSvC`-TaJ(GuHzkZP2lzezTi&@G52AzKiLpcEp$E*s1$QEpQ2_!x`e z$jTSiF(`^z9u7&aE1Tkp5!8z0jVrhF#kME5m)!?(YjzqCW2b>>3Z=73H^@cGOZ?0&2t= zi2l>rhfTMs?K&%<&q`3r`M6VQ+GJ~l)$%!(<1Oz#^gRKvJR?pA7!FmQQ_yjKh*>(0S zU^gu9pfi@hbm5iBL_JxoyLhohokMDj9{z}v4R-`0X+0i};`CAaS(IXw@J2ZM;q3lw zjbL)Y&GN>F$b@J2#8CuU5$1m+omZElS`>u;5-1oR$r%JBz9A}EKtRH;pQ2~2S@+I# zux*8N_Sw7Yt3CaT?R@lzYaIY5l=aWDo%sZ7ECNqr#f#AyLHJie1NFsgisss(!r5o? z)vDjlDR?iEv6c6o6&h@#K6{QHw=LAqN8rSx%C^n z&_)WhSA8(mR@GJ<2sJAfz-h|squu!-2!}E1b=x1TmGvrSt-1ZvF7MKYEX^A|Pb$Mm+r{(>O zz z#bc9b77WlWL9ninKr(1&Cgbfxnq>ws&R(xdwy55P07l@!Y7JzLoTO_4=7zv@0NtvG z#0Rz&Jv604!Xc1i`FGuqWm1hm;2)mlA;;Y$_VzyEeDUbNTd>9Y>xJN@tk$T+SUn+E z0G4?Ygm_&MAZYrPPYH9jaeiCur&~zD&s$Z&g}qW6GlVX8e;!arsm;DU4=qHA-9F1) zXunh(A3jol-PmF|uj+^+0A}wKrYwr5`3Be_C_DBjl{n;u!42CjB$P+W45j-!Ro5aS z5XHb`{U%2CqUc$%vEXz7bbk622p7;0uzBO%2#O@K4kCG~YwP|UTn*U)vxLHkyc4U3 zeHo`_Z{MiMy;I=Ukd9sbbAeX#xyj#iI`sn% z7X7NMgX}gEKwy)urukMYJeQ|~&7Df(okRWO!D0r)QEq;jv0uKG4kd&BA!xp>>!QEzcEL7ASR(bby{!(k<7sUBhir%N}diZ92f1V83 zLGE_19)^@gS-Dp{8pXU?a|R@3Fpp8t&S!vZ=U}UCDimnV)Sbv)?@SDdDJXm$@Y+b& zUCCiVmgze~R852^-;d;&pjPLK;$1^V<+Zp>Mqp9WM+aAsZEba}X$`a*rGM9*GJPKM zWF#`HqU{E>lnb2cnYS zdU-NXvFk444ptX+u7mGbw_u6DYxH1NWya3{tKo5C^IQ_nqeN5Gc|R@Z zq;cr@tC580?<<$Z>V3}k2Lq<$%O(GoX1)0cFPoR|ElX_*!i`m2q5`@yJUY-J(L;XY zEn%aP8;&PrS!W+kjD-U8bXUv|!8d55-(dZz-wT8F+9_mblC`SyB!!nKel7SB|37iTlmVtR>taHcOhIgk+b zxWYHT9we|LI1a?7^|Ahlx$w?={~P-xvr#C%SRsBUi2I5_Ul#g@GLHHAaxa6sP)xjb zt#NX%GIo|gBtSd})gd!SbL>0h4nhH$5*{E;;4}dOD0h?mfPbSxdclUO`3^D};0%7Y zOL}P@H6bZTR3PZW>GpS21aUYmaw*ZkUbvsLNf(ID|GvapQncv{>~%b$lzk4CGe*SI5)oDwVbx;;-lG=cDg}Y%O=HF-h*)lWqP1+9p6i(+ zPmlB}5klkn7;nUw(+|0BkTSz#2tw$-ifp{hAsmaec5N@5NwF=LNppqfNJzZyZe5W_ zo43zGG?L(osyQ^AjG~oW8|tzN>nqaaR(Rf8^$gt&0Oya9x)IZsMIXD4?26^UNUK-# z;a{55VHZ>B6y#rZdv@Fba=rBD^ykCP91&BUxf<%lcI3VLaqVXJdFxJBm;GQqSUBi{ zBCs7^crBxtH?*H}v-u!_m2(C7hr`(*=Iu~+Z_@c0&tb3mk<6#2ZQeylxEC~JuCULa zSR|aG_(1oqr$F71l?9trul>G}EXfKe{N`lc-!}h&;?_siA*lpla?ia;VDH$P^ihOI z40=1P?%kH7mJ<6bw0KJjM_$$_9|#vf$UfJwJxk4b03-dvz;l`ms;0li+h|EbW&zpd_w1RsW?!IP>LG{oBpM4@>LRP93*woP z!f+3f6Ee)_Oy-D-OW`Z?bk2JF+=G>R+80=HT@`Cy#i&85bx+pw~M;5EqO7gY80`@KvSaZu6vic}rhK7Cyf7 zis&9om*_610}UWtLcO`%Zc{T(vrtF3_4}?t2;L!bU|Hu%$mX<{k?|VP2bwUM1mxoN zg=plOdpzIwZ~NWfS!JQ$U400}`@p(FRfV#(f4Dq8cz@v!Z3x>5!t5ViO8m%Qd$k3Z zHQnfrx%}sa6m*|=weIn%9UELwT0q!^rCdjtij_0MFzr5RH-EPah$PF<4;-B)Yr9`c ze>r4lrt#-mbA%n|SJIyp9ZT@A%Hn`A?i6x3Oh8+dm?I8*RKb~$;aw3^F7-~2;HNRi z4B5RTim{vlJid`=4T&Z6GFR^NcNc*ZfsZhpc;RN0%`7Wnri18%ySq1Uj9KQel5Nc6 zOkP9jCXr8~f~031$MV$Yo7yl9CczsWoVz1)*(+D~x=?r^alK*CPfOU>q3#^#17O8e zp6GZX6u$*BF)@D=wVXCc^8K{=$=hB;>^>JRA`RsxReTmO2ypZvrfOFIsH%1DL z9#IpB%i&By5vJE^<8pP(H(F!(KV->y@1Wr|^xQip-ve`d%?s%QpIL_5&S%#44%N~Z z9AadePMuFvG|4ZD8S!CbH=mnWo5$LE^ z1zJGVP#5jcD57G5qQIipFY5ayX9AaIJBTRF1L1j{L*{JIh3T5B7W6h|c9*sCz9`&b zc4EI=(5S};(L2y>5n@Un&LXcNGQmEVujQFLV+dduH>9YVN;(#ABn}+6%tC0MZ ziR8f7oR~jc3jrmv4VdC?#bHd}piI<`>%?| z(fO9SulHp=Q+06|#CGX+Fz0uX^=vSPD7 z&_Chn1#Z}s+WHE(Rz>j0x%1Mg6pHWkx8LpE%%h-qf(_sA`_0%xJe^uYHxgbuCnV42 zSXnV~)g=FS+!3trF!Xkd40@_PsX_9(B5go-_B$~E97zQV!ad->^2oFqj>F(PUEa-GSDntls~iJ4Pv4;9mHaTcghjYSzborAQSX=APWmUag= zwrjL=s{E651lSk>T8O+dI2^#ir4e$aV5Z9<$w;%2g9G3fTy}ugNd>oV0))i3Do_AH z14*Jz<%d_;>s$(YQ_X}T0>V3VBGmgayWcA$=!^({KU9N@DJ9|*ZH&Z8x9-jK>=EPx zfhS|SejJWPBL+jQW462?e#o+k>-vFWAjGm;@O(LQ*8eEjL4 zBPhMs#encrL$wzRHk46LwljYZ%8s=p%o_3z&^xBah-Uo0_8a1}PZgwP)S(=u`DJSB zJ6aY=trklG?Eg#R3cOlX`ayeGOmb)Wb}z&T`Vf?DcpR3$ZNDqKuSCXWl7+(Y)m#Z6 ztWC>lat~L7WRmQNRCOVD+TkO>8F;>%mjm^jwS)!g@EOcE`7~;0As&bnfg}4p!}cBi zn`0(p+asCAZC}5d3@Qd8^3IW}RtcgyQU2Ap- z>xammYDC)LnP}x_UIk*>!K6qfyw@$XGt-k4#BG8bj6;OOV?Jgb6vOacq&FKl>rJ*9 zU&>)OIVmZu)0#kM-)6sKTM2hIAfdm{Q+s~y7tt-p@13P}eu8q*v%V@!h^K~Ro{tb| zrj}aSFEamFF2h#ds3ARS4F(8*j>A2&7WMmeAn7qqjLdf)LY`N~>wCm2zdC}yWpH5R z_Wc7nC1~~XG|>7j$rTh-Vsb`M{PF9+Y>?;(l@?Rr2|&)J3kW;tk2HD*MS~+W6>YUS zoAcmM)omy$<_d!xpBXNtGj#4F@qt|_Rp^mxpswFw1U&vx<-M^Nz{L1B=7G@$7`AL|gFlCJ;zZlr?fd5oTjx_6m0rLX4k~-}zXtaBqjUl&g_Uo>5jdCF^I26V3oIU;h zooLYRy9V$y^`|c;num&07?35wJFwlfL(jRrtl&FD>}q$P`6d3u%7Y^Nv%kE`aL@uP zyzB)`c6n_(FuA3HkX$AxQ9T-pQvulV?>0Ns5zTRp0)KotdT*0g5q7sH1e+Y5BLc@M z*o%7Kcz;XMQ=JbkOJbX=*3y&Y@Vfy1+XP(%oIfxXgc=c#x2z#n#Z1Ho&Sv$z6$lE} zkF#S{4t3@#gddEnd(a~2>>EtGBJbugH!7lM4n3>guEHWRma+Sh!>}oRGatw>q7tK# zwL+2;Fl^|r*sz2iu4Jmt=o@?xZ*Jtq5aEy*Q=O6~@VNb+WqG`1@{AWiNu?UmcKBN% zm@$g%L`HiL_qT|dAr4&<`!VhE!K5LQTG<;Kn^$9>X>8uY-g$W~kcpE_!GmZGsl9RN zHM!6BCdb_vbiQG06laT&mcg-QUung z_xL^*+FLey`#sFpZ7+1H&^U>rBA6AQ8f-FZ39emek^<~k;Pr;_Z-34`Xl0uh0LCJi z5X?!p!@hvOt|`K(@btVam*1fO(O^5hPy14p`RNFuIwX85PQtc(N(Rz?%d0$?Wrb35 zExNxLnVJ2Xk9A*&Ku7mG_(G9_?miF4(+!x=-0b~>OEqA7L4#)SE6$0?ZD7|mMdBHS zNa*d37(q38Zk&MPQOIgsLdS2nE$m|JAyagMsS$Cg0S3DR3JmRA2rzFmJ)`=f>SYP}saQ44;lTH$Kvhx%iK-1$jJLfYECDagXO&d8 zz;MQ%;SQcM(E*|Cb%+emG2zn`Q%7>Lc7TRcCy?^1e$lpgQ=@V3Nh#urDlHd{S zlOEo-_MCfJK)k);@NEh$qH}(PDB?x2ry3Lq;_IH5;=3|Zy{=JjxJBo^f7SPa^kI^? z-lxE{q0NUtO0eCo@4!bxv`_!~-9*<%mbRdxjL3k90V??swvYXxjq`lW)8NGwLZv{* zgI_AYpuJ~#A8`3t^`Z6ZOE%mUw-6?cAC_s`k0*cQ`DjzFJiErI5jwOwJmDpi$YP1UNsqTnvh?7QHOut^$Ffr!$m<@BdWVPkx528F2P@dsD8nbmsOY9q}rZ_ z9|`^~m?@j1Ye2`WgJ}zC)wPAmY5flN19v^tx3s4~JVhDEN-A@xdsXdnwrK0&mf0;J z;l9z&h(}9K7Fi$(9mZMyay^`G*?V!rOmC4>3h@Vwy=nSQ``_3eLB|ssa^0(e>52cy zpW?U}xJcP7Je>S}F~vtYZfR_RM9RXR7=|NHtHi$0 z;@aP3wqq7JFg&MrXds{GkZ^1itFhg*1gzC#%vN2+t2MWto)uqH|2#E3aKxUvnOw~0}} zyE_F-(D*Uj;sav2oL6pbI9;&ggd|#=b+xWGPS5Xexar`Z1R?UF5eyM{UroY{vcF%i z&c0>=2i^w#dytIkuhnx)b#xtSZHAc{SyktWo8c8tGBaqmo;KV|kFIo0fVI?18slNz zCL5T)(e0Er2x^Jbpo9B}sy`i#O|DElC8Qcr(>8ES@p!Tj?JJu2d8m_gRFrf3yH zFD zU@i3~mw8_^E5DR+cJvlJ?xUzxhgJ{iUg68MAkdZ$27+04E|P!Gzh^!eE6HVZm;pM7 z-wxTcl(wI9*opim_@5X9ax8N>*nzfiJsZ!{>w5wiCvZJ_n6s&=IStI;;1u?psXvUg zn;iCF!->UxLDjPk^Q$?B0Gr)gXMZVTPyR^fc7x_Z>vJaTgZQ|aKAAWG#GtM#)!IaL2Mxe%L zqUdNV8aGeS6pBT}Q`%^3beQj-4?1bSkOP#~y>&kZT*1hQp(KWIAm6q4zS^f@72%9a zx_h0JaCB6c6ZqjbU-Y1L=${!MXAi*_HTK7|X%OAi_j*i=5EBHzTC~~Y0Pll*UxgEA zqNhgrdP*Mk>9p`4ccO+vU3%y#eYDmAJ7Jsd_gAj)`39^Wsm@iPACTTqRUeOGK9#LK z#BYdB^U>Kl`k6KeCswoA#H8U)9rPFfPER9Z^e}6ZsUg=0tmD-^-tRUOb@K&o@jd3dYZ4k+9miLu7Yh>&j-jCXidPWj0qz>V#Si=r z^}|i2&+G4?z|akN6{xm~`a$Sm5#xkV?qUv`sBCnJ1RC08qvfC33&|(f^`7^UT`Z_8 z>9QZ1CL`WZ0XvvmLIw%lrf$rUeBQ6U(&_Q;{9OLicnD`AfP`<|D7I&1^cT9m_G$jM zf6Q6+?$`x5ms{_4-S6;HR$>va9BIh4MH1O?+s_B()inHWdTGimAbW&0ak9*c@2}n7 zA;Kf-(t1>zb$?uk+r8SD9PykzXTxvR!DRf{x()4*5Nz-@gB*K)(PmB5*A2NT=+arOI|WIa^cyhiasY~k}@+s%L|O8Z*lRhyzdO! zOTqcz{aU}GJJ~|&k?!&%49}t{MqogrcqRs6kY${cplx`eg+FfWDcKv;imS?^`qy1H@4V9qx

>sj*1SA^)d30h= zl>j*`5eSV^yNIjEs+_Ce8V4~O&3mHc+x5J0M;@L|&0jp)Gi0!yb5-9{Y>HhB&0wkG z^Gi%GM=&3Y5&SFs62=Py-pWCjPfu`ql0fjlSGUhh>V_FMK~r8J&{{`&#udAFdOs_h;*u#ku| zM@fRUd}uVz&XUZ!Uc^aGkag2$*%|ErE>^xpQV_}A1&yU`_YIM zHld|)=;@Z@Z5=)@vaq@b6L$}8cE~}q_rM0_uFk`t7Imer2!%t$eEUX9b7dct!a=>{ zw$+p%6a4g75>%9IPlmlB;}1My%qp$aCPh21#dBvZr?T*e9p%2=A+c{WK zMA^A8yyT6wbF0uVxErtLV-x$QC!P42BG)Da$$rCT1O_JSeUxm1sQ=-PIub*SZ4I?%yK;Zb-oKe2B(1LM#NKhJ)LB> zaH-&@9=Ir31P=(U?)rKM^t4&j9UUGAYwlv7n&{S{CIfQ1I^+^TJs6H-Y8aOoAQVZ7 z=Si68??rKL=|hrK-)j}Jzg!dlGN60p#`t2^=`gokkKO6>To>xYAVyWzRsM}U6$W{P zsAt&^!F2bdR7vQ?u}z*=WBoK!5@E6XM#=oXdtr#T{YeRU_AXAC(lxLk_qgnh1l0wv zVDBSg|3+cC;}==}i(C7w5D9g0AGN=!wouEd6 zaVy|Thw1^tZ?-n<`VR0|V|o8|SF0qyB-U=?@*1MZD{8@P>?TjH+aviX$NngT*fjfJta{TKtM z)dyP$yBu_O@qaDpkz%GdS&bZ@r_;IQ_k=5&y$J;Md0_FJz3*m22v$_%Cn;-iB`M)P zuPT3E5+6=j8@IC$_I+GvKFl{FSXD7``j$V=(K=MKT!j19I;vB-FR9&Is$a4t-D>%D zitis;25vO;8RWPpup*;sZ!}y|MsQ~RT z^>#?NEuKFWihAC7@1wfPCdXsQzEf$N5K{p-1P7s=nm3{PMy&mAUjaVpYz`hJBtc2z zlQ9!Wd1^)plu_V84~0?P_V=lTNjshQ!%y)wnh?K(=R9HKpA(6ahio2E<6S8-8pQqJ9OfCa{2c7jL zqELbB2zk-$tL0HNU;FMH>8z2idaX68nB$ov(;Z6DXeF72tE2vG$QP}8T5p|m3E%rM zK770`h$)WA=;yy?lYmJ~)@#ceC{lhpJ)$Ra(ZvHbB#63U3ShtNN5^!afJnGQ(m?p# zpz;~mq%zhVSvm@G*H`mIjZ4d^i>kQZf_9qiF@ktqAf{87Pj`qzz}yR^GF2a1QE?%d z>&;n>UC6dsxAic5oC#E_l$_57X(6$|6m5irASpu0tHIVRm||=bISB41;I*^6q*b(| zcV=|O+qboxg-f4Lq?vUuY1#$9(I}9v06|j4`YtRYkPuz%_V-*qJ*N$uT41=Jv#9#L z2S|sDhf5go5Y|bK1TDR|hF%_C%Q8R8L+Gve9DYW46H%aQD+0?sm*eJsE3ORHLvO3) zIzHd$UD~lOV=U{JVMJG~Poq89&u91+Ia6*ToT^bIip%b6W+Vrdazt85zU~7gE?ZO% zqc5V6I@a4yLLVbj*DMb_ZZr!yB}Qm+mht>rkL(y+Ycc^!PSj+>0Dtb0LE?aiPRQou zwJ6fs9^8fg44@Ej3>8Yt*nXmAgWr9kOBKxGsb|t~d z6Ka&wh$6rj-g{x;_7@Q*!v2oY8uXuk|LT*tQ7m|E_jb(R{V&&A4RRm;*}Sk!v&Ng> zC+O_o#P7cE95XvRjDmd_LB%L212_1+8%g<6{?z+Q0O7~b!&m1@z%t+)q5mJ9yMw4h=JxIBHAzv!ptnhCY*-SFvsUO?V zXTPp`GF1Q=obxId56=u?;77!w-oaqYa|h+2nD}gOYvNJ1Zx0wYqcjwd`rLb&fwXZ9 zJ;{@BSrM@*ec<^va84xgREIqF1r;%?-R1#30)&{%zUhK8S%|&(_1ovc8?trlTZsS< zzTK&lW@u3%yK@m;4MjdVk@;nNhJ&nT`dx-7P}vODZD7Csb`KkILBG|waXtV&*r+jB zyuIaIAF=_KD1@rwDMZcp1$q@mCK00Hut?&|hDVyoZ=THiU5O(eA0fbLC$m7dZPLCl z+#MGAx4$YGb!U;Qd;1(545(DJOekdjAno!a_j? z!TGlh(&_Jy9jO9sxX9GL{MfI6{fQpKnA3u-iemtZ0zm4Tyn;hR%79zEHMiGh0sc{m zkKA%6Xy(NOA0X;)>A<>1W;|%N&+!pTP&=?oJI}w3x%2}q6Zlc$9k<#6QI-2v#wM`R z%_8QxFDI#g&!1WNy%eLPUn1vZdbR&M!fnFk`uEkrJr{?KOcK_Cv_gOt@zNcN5iN^zxDI9ImgfZ zi`ATPec=rck-JKcNiMP`Va>=P4I{F^XgP0>c+0!B*O_92}Aa>&Th zPqRu3P7mJE{t zn+2erFjd_O>6>?M{`8|^Fl0s5ARc+fhQe+s3Vrd@g_sn9Ju3>`T2TCcAwoH+r%QhK z19Tb?gMo-B2C)QqSh3DN!gzNu9rn89q2Yc38~AC4UHsF=vUmlB!ht9R;tB>c4XhD1 zQt(bE5A!=|sG$&x10H&2bGW%kh4|jM;B>gXyz^imLqGvu?6icX9k<;YGRs_AQ*0lscYm41v(349}NG4Tjyf~AP8sq*|M*4&XNPFNY zJX7HumMP(7SAtHeW_|?i?k;BKWVL?7Yp*&_88r#Fy(x&iUeLdLeedDR;(jf@<9NbC8q<~PWMze$A? z@#dC7RoHQRVZ}Pp<8kJ84^yEssp0p!aD8uY|InUs+RqK&UDz$II?2b%!-lWJ74~C3 z<-iS1M4&wM<70gU^>qemqPV|X(s3jBpAguHiWpYpUDj1o>cGKa;;)MnZ2Ck#UI<3f``Hh|Kiq*T^h{lI3T&-K)TlI!Q6|dQ6+4wp~!8F zHCg{d+_EzZcHOqNu?JwhG2UsohZ#htzw+@qaCBK*Ub>{)M1ZkGs1v~EUmKQqFo&Cg zt&LUC(8rlU;3x8sklEgXQ)1z>ZyC#J!HUYxt3Izuxgc#%qlUW?YDf?(?_Tu@Xbb(t zufbgWVy#|SoaZ8Ek$SQJMgqQ7_L6u05tJi`qY=!&`mX5vEbATn#%ekMDA zpD~C8zRGFJBBQ>KI4r5K*)Ic}&Z&%-y%&*styhWG2jV?2xqrA&% zo2{9>RGZpBS*)?ZO)2CLM8zf2=Sj9NFh}n-@}ld*#q?}VGm4LjytweheO|pppqdZ( zh!$>tB&5D1jII~oDUJ?eKhZF1#iSKzpk3dAUzhy=d*7KH+aiE{n>G(=jz@Rg2jlx+}hWu zT`y4qNlu4^z$&gaHXRYjb=ZW%WY=_V*`7T&+>XzE7`7I!6C{w!y~3Ww#(=vHkCTgq z)T`GAXM7FRRicNiZp=8(&dc{5AITtBeC?xxDO}x-h+C*XFGt zDlH&8=>#xWY<%VuEZjqv+Hu7`1NpOlDXN#O=wj&~yJ{guSuysy=r8mAN2zieDg{s0 zT=U4C-_-*PIg7SqK#dA(j4aMKr`!X($a}$zQAXGRSxwfo@Bp7dm&TYriz$%Uy`gkM ztR+*Y$Qp>mR}I68{qO_nWoyc%hS(`&v0sovU6uRCPcI>JDi~^Py!!8&)rf0aZEjauhQ9El(-I;7@yTJK zZ;g0)o?TtYbhlJW+E0E%h~Tg8ujzy4gQ*sJRLR3-dG8kzl8Dx&nCue<+eg>Dvmao! zVPvlz6x<^Jeg3+VO;f76xXO0|TzFlela{@@Kin!e?APW~5Ycl%eh-^gapv>ESDZnn z`#`VqJdK6}RyGnC-SK5Zg4bSgJZYhhhkJi!F&hE4nGW9cKwhionR-)Lap{#UXsU85 zO+#wkod)g6PJg~CYDm^P(c43GA7LG&lc%NX4KK(|>v-A0NKA>3FRBO1>k>N%9BTxY zF930#&m~;;8m!3DN~G*@z99@+Wn59xL{(nDZi;f05!fRr_*WQ}a`?v(UF1Gr3!zM1 z>S*rkd4wd=ux=2;vXC=_bzQsOA@V(ndwhPp?mPW7v5%|=GZ{t? zywXOM{RQ^{gx>l`AX_Tlsz6dJcl<3ob^Wpo0K!N6#puH)rkYy z5we*Jb}Jr^2vTH)+sIj$mN`T2eS3)#!5ipg=O41R}e0Y5w!Jwa&2I2hdXRQ$*J0GG|LT^(>(@v zK|Quxz|cR(o0J^MUXV3{__!AYzdtE(ZG>Bnd)-QOZje@xLLEFL(x8gwKdtj141bCnsmsx=I z+-Qr@wSE8)N5qUgb)B0Efh8Fc{_4H4c?3%LR*3FEe?)F?<6K|${^~tl13zj;3U@`N zFF1E&Ked3}WY>>}8ihQNUk|V&bm6xY=)eM>+_ZB^fUSI<(;o3Pc#sDYkcms`!XU9Q zPWQlCn~w0Bw){iDoyIY(MMFV$D#_Crf_G_K1R|XRmN(lk7x&v80V+bZ5J+SJ837my zvOB>i`SgaPBs9L_*~{qc&(`|Uw~{>&hRO?djDyKJGu`q0sD77&I!EbQ>+@wrKYYUM zWOoFzCySbUl#_}2S;ihr1#3Xq>r{s?2I583p<)@<=yEZ@Gvw~vBc$Z_9KZ%fe-Qm3 z1r0QM-%-r>YnoI+KX>O2+WPn5NL-{?P7&c#0zG&Y3NmJvl&eV-bQv@#1*YFycQeoOHcO9}ECc)P<9R`AT9Iv@70W^qTu8IHehfNKnEVP0h(M}zMp ztYn1a41&R(NzwTIi0En);Dsf;k(4xqL7xL=9$$)f3(0kw_uFG+=b#VorAS%gttMx+ zs>BahGX;g7)<~kaLZQ!wIkG(N$N6h`w6>)~X01CCFxc}GJ~*G={^m|<0X)5el&TPL z@jq4KOI&^F$2aN03|0^uLF5lG49&=1lGnVr-pq5_CGhLh;uUI~Mqwjv%FLU9rlc$) zj2l%?-?Scqa-Y0jFl0xHrQJw)j&sZZ0w8DN?Pmik0`ZbljAadnQPD#_JxQy3p=%nuH{;GWL977vW9C<7>c_Us~g*T4Y=Ss9W-|4t4|8k3P zzjZ45soO;deh!TapRH?ImP9`-n!yp$U)Y zcf_C_rvTt0d8GQHerM23pz+@qx0xv>mx8N`i|VZEiot$;Zrs;^P<>LShsM9}q7iT3 zPVU^7VfaGqb+5m?(20CS#Nl}Xle3%q7@*8$ctNxG{g%r8WT{B6@DwP%dL$j=5*EKG zlqdum%-@CP@azrhJ_kDZW{;nUA}!W zy!+mnS41^7bj4r-Gx&ywqd?6!Pc$SEZ6C^^aCTzsfBkDkReKfNi+jmL-&kp*xLqDxxP zfT`lXK|e?~aS}zvqP-`%ojg4f;IA9)u1!qHwAGo0s}Mm8XgWSc_Q_Fb13u6VoX)2R zbf5Jd!;Z`y9j6dIp~39Rf^yzP^hdxFN6{)R>;5T}v#WfR_5NHFWPf3Sa7*UXuu0(V z#}2?e7^|}}i@IaVlww~4u6YUiJRe@t#TYQdqKRGsK}Xwvi7%eD+HxuXnlh3-c6ER# zp>o2!EeCbkm!hVHTNcCu(cnc#?cM^VMYC59q?(e7b?l&Y<~gvs8p&4?VbZiM=E+}l z&ePYU;o^c*^*xQ$yZtZ&B9w!Ao#x*|`k+k#=+g^jl#QqHu^l)t?SdJ3BoIiTElE-S zHRF>ouHeC>1kMEFw_u>xmMy;$#tpG@_`c!`xkfK7bft4*t7{X;bG8HEvQ5!s%LK4X zpk=N^3Q8%4Y|uEm#VG-=z&uWylWz;MWKa4fu^66z0}c#=g7##;t;V~x)Xvv*K4r>- z=#FnA{Sc43-6Tyl#u^ejbdi{7*Q1U-sW3>i{31RB@nnDqrswkUIW^zCsF3qcdYhle z8|*c)?1OdfFssdm{r)(S(CM<%f8#XG zuSBdxm4RV;eN@T0j3()S9E*! zU`8Q=6RbjPY76<@Chdo4V$k0NCxRi4p1ZHg8!`zE(yjdA1b4%@wH!{PKWV~)mtoT7`E!wpsqIiJW!=nqf#xKu0z3H60;avH}Btda4 zVIJ1=dFgRK3?F=U0)8^`w$jhOhpYM&om=PIUeY_!fKTc03cyrG- z>z;RwQ4~Ef-Rmg9bTOgVrMm>AnbeA>62PkT61+Ex&f(RnKR z;27xOZmVt`8b*tQ%075nY-*z+8F_fDLH0fJa3V-zIPl{G(lIhN3eCOy&J)JKU)Swr zhrOyrt6j#xV;c&<@Y;{=|vffSwHW0Np!w_uf1;? z0hI><0&ly=K%X%LDC~eiy6qUFe{&wyMFsz6Q8zOcg7vB)jy zG?Ra<)-l0Y3)>>Y+OdM*gTCfTg1kt5g3F$5uEL-hxd6tO*R3}jC=7liqQmr{N-8VR zpY&CoOA@R$}7Xrb7K8MD+H>7|?Am@uO7Bm=ogzKW-hfb@|toZ$MDKyGS`{|yI zkM;Nb;4Q>uy~09Q8yCbh;7g5q zDG#0sU7~B81;#jyWf#84*Y7SOzX(zYRgXV!yGoC=iDX4kK*qVb*A+q%GsW)FzOY<+ zpnE(mP_r2ivKwP^a@nXoPGBm=fQ{q^(LquN??1WV10kY2U<31`Sx6FO zt9A~MvtGk+ChIb`2G-e-0si`-9S?6!8LBw?i_uAxflmh3wWw6k#_&MLo+W8x`Q*b; z?3lmIZAM^7)%MCAj^9>}DylaMU_%@kLww}GS*IEOvA<3NFj3U=VoTfr2Cq^KTdsE| z-864vNFBwxln$2b&*}4H{~kD7x~4ZNo6+zu>GE9vLTYbha z_CbAdc4p`3ekEJtPQ&h`>@GF6!;wAH}%%FFOG9!~rKK%)@&ezr5Rh3z4D(t=2S`mNj zj=KUn2()z6Lcy3wm+MqD(YZlSz=&dKT4UMXQZyKw!t_m;)4>BWS#Pp#q?`(3pA>6g zca~+Jti4DxNPFP3q<6t>cYr241=oH3d#h*=KRJd*P~(2h49GwVC!rJq>9?Id=YiBw zR%phT&isx5VkwY9avUu2`VAdQRy#5u1h6|>rNve5^cAvGc;|Q+E*|I=VVDj4SU#$} za9zYLZBL}I=jKHKG4#Q~oNbJd2V8Ze@|>C`-@6pcstpMDI5JG8q?3q@1{HxqGugM3~k@j^6N@hR{!DT28&zIBGiZ?=NWpEZ@}`^oDn1jk6Ip_rSV=wrE?b%KrkWHM|Xzuu_c}*BT}!DK_op~5ln%bQK}Cq+6BZT zRhZNBgYi0`CxWkCJ$y=~1DC=1I-q}!R16GKu@E7M56`-g=F0*Lf zsqs#=l^<}0mG1X#!xL4asv)Yk$^1;_@Y?Quml(MS?_yhJx7$oUVcOM1$w5k=d*qO0 z`g?Xqu;;N2a4+k=AefGw76Dzi_8lE=E>jiWxx4{p=@;wXUN61PN9ah2@5hH_Oh6S` zoRlIOibr0W`-hmCzl=U-g!PruBi;2WCCHz%q~^!=^YVIHWz+lq5!uJ*+Q@(Kz7-dW z1m?CAph3!~A2&s}$heI4r%f)ZTimM34ELZ7vb8Zo_+x!scK!t$nVg*WFm3E zS3hMQdYbO~XqfHg%f2Zr7SP+&gh~ zgpTXoki*(=|9(N;``n`y+@-?R8U@~yPDbytDRC}QJ|v_36a`|Kj|A)!Ug3mwYx9}Y zn**tFzvp1H!z<`y^8L_W)}vL!M{<8PUrwJJn@yY7&$+lBb})N^V@l0)zHH$67<)jM z8e)zv$g4TgpsTF4KP%@4#*hY~2e)`%>C=~oqYC#a`kUT)pBxdC3mkt}XJ}&~UXFLX z0~ny~poQ2>8-0H>zWV~O-aIkY%NvNbG-7S!;YJJJu3!)0dO?LvB1oaUitVU#udKooNHSs-2@V%Kc{^Oo`&7`LF@34b#BxbAmduHP*3GhKO(=yfFrB|V= z;T)gjfE|H@)4kan5fbPKKU8R$gVvytX^D8QPSpHU^m>NDYwO;C#j~3JRQ+mzdu%UV zL6pYSu^?eePWao6G0L@#5q2$mTrpGaBM5@*6o7O6*+^26?jxMv9Kac#fcr<9{jF%S zM4C?wgAf+Xon1g_D)<$)lem8(uQ;M)y_BK<*HLV*FoA4)!PWpFsI|7AyXXukcPsUw zfU&T*waCE)I)Tmiqo?(BspAqgsMKrw7OA#hiP6j~%3R4Q+dMJ3+xpjP7r)ctjoMv0 z5rh-W@fn|v+gt5@GJD0*`w6MGzg8e9D-s<AAyrTc5ykP0n5@Ee9kAn*^- zsb3&1s@oR^m|#6$&g0Q>z>lED#(mRhO@4Jdmgm=vKH{AJj*T!-OpyR_sHvZb9ul_0 zO-;^OCKX2Usi&)X{guYEj)9?FPV`)!)gzMO)G~U5oeCB$?0^*R;Z;_FmT8bzvdeN1 z#Tv{{z|PQihMa@f?0L?tUww49fPd>cWGI#9Ci*UxJYBG9j21v4gL4aS%4=AmR_H;(heE z3X{g|-ixN5OT*RDZ-^V+z=aANCx^2vUcua~P zlV@*udl!^eZU;m1g?pY(6?Ftw46}NOL8s1vaHNM79tWuqf0>K5ks)o2a>V-^M;u{_YMPCNUDL*eGuj&KzAu(Q9!?=LqXkU6jS z-N;ourW+49tz`-(#l3p#*5}#NKl(VabRzb8Z1Q@2(tQ_DSw7^J25G5k{qcw^xE+GF zAGnzkC$?$uPQ(;he#xHHFHrdh)lL>oimooK28T<55SN?esPI~1EQRz;os+2cVobJ6 z|3WG1dNk2`PR5iIGsURJWqZ0?id@}Fm;_p195!TkaVq}9b>FBt^EFpiy%xxi#y)O* z#tLQRPUrb&fU(HA6Ve5B^K}i6Z{?zYG3>2#Z@9i9$}&$B79a^x&1B6wdC%0r8JyZN?aI-#={c)yp zs?qin+^tJ)N1W%2rrO-!I{Y1Bl%poTuL{DY`_H)K;JJ|U`&-Sl2}URLqCF*Z8TSfZ zNQmUgMn-=Kwsjr%t`r*LaY1%BM?W@|)6_mLzyUtbAZ(U6dEi)l=d;%kTrH#zmXh<) z#fo@G%wblW!*7A(_IHDETh#TQC9G&!Pnuq7bp>h8Rshru*xM2xC==Yb-#y0|bij8!qI)e-)P%0HC``QHq*~%EG2^=)|O-{gKa>aU?_^cUa?<+$%jB=-%@ zWsFpxK&tx+`Fvay`iYXxPlie|T0ccJVtjvP7vDh}dqq8@eHj}Xe``98eplEIXH)#^Kh3aAS}U!L6= zLG8pAS*Uz6>+>|}30?b4SmE*;Z@M{Jx6wqC5_@VMS$LukaL3Ery8ij8I~ea8{uUs0 zb-U~bD|`rzz(*g%0}wM6pGvCBsC}_5Pb%S{gu>XHA0Fe2m!DEYYJqxWn0K^Zz&yT) zba~b6uxTOF{*4Xfrd~)k*k_dm2~_YAz+_3+5RT0{ zfvWUSqFF1s>NfhL@a^MkSIYP|-Q9F$XvH7&1DG?U_&xC>CW-3U!=A$_ zQ82O5_z}f9?fG^ZlDtm4Wu4ORg-}{yv!QRdSi%a9Du_`kB_a;mfj=4oCqZj{0o(?(w|3@EgAHEbU0%Z1w^l zNDAeJ%h>_{3>2qK5$TAiBMC-qd{^y0(}EfgvVp4#sFdR?I$i9DWXAA<6vhxT7S*db zquMv8SL>gz^t)BA8$uBp*=%`erV#N2-L{yFyv74cfi&m7^5+B>2`-*962=Oh1-@$9 z8xHvqc+O!&ILr2>9KbALrxP_-@Arfb?x?AVd83>UwThx1j<1Vve*%GrV1J9l1ZR`` zfJf~?S`o*VBA&x`xSz?VINp#5r3YYekx+~*KOYZvwn}3i5J&L4XqEZ)f2o=4o}UHZ zSN5Dj3Bi9X;!vx5OxyW@#2bx8mYi*HjbeN{Hxuf8_#<*4*=oY=e2RBh8bUlNbv#+7 zeWzc#*-P5+Zb`}U-rxQ#(l6O>&$jFUlKR+RDPwIL#9#owLn5H5UVnKWETdU&2I6vn z2+z3isli0JGE)F}b8VX4%f^t+9gW`9_Ar#Cbm=SV?I(EkO@e!W5ZS^d0YeESwdQ}`QOFB8F%~+WvA`24Qql~1a z4>|wl<6?v8*yF77|saJGm02~}3aOUmJWuRfYW@2U0<5$#sfGhb9>~8#nlKg|7FI2 z9x3Ql>AV>1L?@ZN5I7IOOZCOiZaxq&TC~Z{4Hg!h%nhGzB|IMxTI|O-;jO##W0l&> zu`a?9Ej8x*?FG?VwFOaNAK###pj!N{1Wo`aVpI?+TU7Xmxlxz|avuN|vtOUj`G=P6 z8$pj6CX7T)JdQ?%#Y%a&VU`#r5J!`|(B}bX*z|PR`!pY^x7G=n4TdMy;zmEm?Sm`? zi2a{QeFl?h-CsWMnisy}IB{})?kDM)c`DO(eIVd*t=?+NVDu-6%-M3(O^vbcjFk?8 z6bBnHzTnTm0g^wOsg&y3pgD!SdFPdFm^!N^q@`ER1hioV#6%K+!ThuiQRJwx#mXjt zb^$Wf###jar2+4xywL&vfb>+)C2Z7lZ_u^+R<+W1Hbp%Z^(5>3fh$%;Lzx{g)~6k1 zHT)I2c>+uev^b@D#gv|i7t0QBBJqhWYj1EYrC?@G7I4u^yI(I~3*zcX3^2l)^x>Ku zr<34q#04nkJ)2LWTY44Ji*;L2_aHOrgrUpDAVGL}+(2ku9UTyrn|;3T z^72_C16cHMZk&^+b@9tZ_MKMp0B&Iyc*RM4TfsH`%UR?`mUv$t1~z#X?4C zRB(DbYmTEcFtt6V(LD1fAhY~PP0W8c5=xP(fJ7X;9KK0NZGg*_iN==Z`0)8ZsG z17XHGspnt4=`P>?QiOEV@#}$l;k~BFBo2odK?oELKK9?|W;) ziiK1J`%wc}yb^Q~o5Hn!fA?M6AdmVBG*Yn#GvG=boA(*7X^;HmFu)NZ{Q>c~-z?v! zD9QP$aW?=vR?qDDYGBIJne%JZ|4zZ z{eBWaOZrSA8Mm-O{~L5M&+8w*8rPWQ$gk7vr!48}p;v+v5PHd-@4n)2|4Shi{0j-t zTJ$s?jl3K~rOpE}KpzD8XK{Cp#g7YUT3&Fr5ny!kFy@~{Vjw?W5QJnzxZa~4idsBa zN1xvzqF>Zfx2uXwA-=hfx99nENKSOlqQkYfoC;UiH)8yHBM*lpCs6z}F3XnHAd=n<{=vim;0_&0ImI6`w2a+AGZ1wGwm zoxY3bmeV!D11$F`eSKsJM|*#vY5dL8FhlQ`D~}1Ge21B)@Xuf=%r+bB6-m%?8-{6! zEPD`PZ+M?L#eTG$lRE*7+3%9D%Ct|TW4ZXBFJcku>-z**Wq0!9&&QT(>q<1Yv_$84 zQ4!TebUub2HYEF`5ub1|z`JK2RD42`Ln1{SzL;PKb)W<)pf=Sw&b5QvV#`WC$G*-z7tVJ6pS)n%!XXs8G>oH-l@P>=qSrd@uLve#kqedH<}<$6jl zvK;%6aT?bQiXQJ7VSDLh=MH@LSP$kc(D~mkihN6qQM?Sr%GK`!a!d)L|F)Yj0!MK1 zWqlt+oih5@^^N4k?pDG-b(p(}3tyS}%UJU_UZ3G(KkxeYx9D?8hOp@D$k!8c1ueM* zUFId8WFMd}&L`yWE1eSxb9fl1;VqB~R(>yovvywhEe0W}E=)O_`<+91R9qh*d2`oV zU~B5BUmEs?dJa;a!y{?clIfjCzHDENABqb`tN79>fC$Qd>cOk|OcB;#?xLO#Kp5=2 zwS0pVKmZn*KMGB*XqT(KAWG(dw$d5%Mm^BmOBVGsHWSG^;1qnL52uGzlMjTSj-4Oz z8Cvo;4odnbIK#XRD$)Rclmrr z>%n9WLO4_a`9UdrIpN)fD-pH=$ z!J^>^IwErm)QcEa#c14@yT4x&XarpX5j?_XZ$1{l5Jp&X%Z7Ga8+0fVu=C#B4=E5{ zkO#lk^G=`pHO-LdkUh4}8{JP>M=_}f$0u$n9g^T0I6Y~8&s_odVf}pA7Y{M_6J;&P zfOM+W86NF=?6l^atOK@j9ne|E`NE5f4d(p{`ld zH781`_-G84b~K%T1f(M#bo(5r(-~O7qZ#A)Wy-;?<}C&xF`+M=A|Vt$q@`BnvRSMA zY_#0WAjNThb)n+6j<4GVLQrEbLwN57(}{CjVBJ}Wn*`yT_<5q|@0AjX{w_M&KLHir zmb{SNq#QJ@rQXuzq#$VGKHGO)1K|mn5RvcE!&J~9O%I8m78Xh!MkyZ56tig^3P2HC z$L_Mt3$*S$;>-OmLzwzVAJUWc1z%E=`FzWMu>D#p`PkRx3T{O=t4!J`>i58}@V=Nd zc46SG>RUAB-{-iGyU>~sr|VUyX)Y^45w<<5zg2StLYoKiSg8e|n0Fhih z_&Va#Hu$L-j=$&S^*YC_1sdxs|6(-4w}TE(q09*<_T?_EVd{h&P^WCk2IHyIYH*o# zG@yq(-*2Ojc|qTJq-aV0K6*Odl3aIL!PfoP<$9LBF9U8N+!m1KPOs626bHf7dVh^D z5#(O`l!K$!N4;~c6_q_C=$2_;;tf7rW_AzP+@g1fIH2<-HT)0;;)Uj0gP(K! zq-mE$xqUMHDh`#aX8pt)3p5hGvL|U`h9Pizjk2vjQV%L^Ex59mvVN%#5OC4qYE!pa z2|{aX(ilW!U~N^6#^O)@?&T_WIg}L1-MJaXS1zT}pMI{6mOyKCY`d!JY^PUJIw4^) z%^0tyMHLZcz@}yY5=izsQWchPLI|C68 z*cl`p6rRKHnu;k$SEsLXf^_xSf#F^ew`1}?14B#XjNu?ezRk!SHLR7J-vTLExP@~?LT<5`EU2$x%Hf+@d-`cdN_$9uBKms%%dd}^*Yb#vr+!8l`J|5TjE9#{X$hhL;M~@vw7&C$1d9YH_%*bPXx!YG zKq9(Z+!EjzoL*+AA~>{X^ASe#_vbLd%tsF6Ty!r;4n(`(dJy@cRan*OF#L(M7WmOm zolxka*`E`{!c~KNMqt)!Hay&DTbsJ{Bd(UVD&b-sZdYb zUpAl!mQo1>z9%IdG5ql=+|0Pcwfbl(xZfFv&hH@YBNO->9%H{fK*O|CpY~gyiF4Z! zn{UXSPGs67pfBq`_)-h)_rR19?{ef`hB6(1wpKuM*%PBCU3|;Pfx7sLEQTPR?fBxZ zSP}*lUsr6JrcC!b(7dzBzKYWO6YZPCZ9BD4_%5A%XU_%AZW=(yoHL;CCl?5?T<9%b z$ZhRwwXZD7rz5|(RXSsD$pb{hqW6Yk5_iPgu%>55$q_LMhV*?p%prkyXtih$D8H!S zL2y|}m;mqb<<=$^xoCljq7(H^z*6z|5?X4UJ{3|~=RCsaW#rotcqn4^o2?>suM5vq zM>6SOD}i>ZN&GpP4i(Mm*5F|vw4tVqI)2OhZ2FKTkK3?L9&xg+T6`~Lvk*TwKHv=T zU|d@r>#+H*p(8L{V&*eMo7e8v0wnBAZp8ElcDDB>3ypaV=K&GkJ2+oTEQiKDJl9 zbU?lM-8>&l0g_A-@Yz^Rz`64)lOKtj8lKl=eEA>$^7>GRD-#|8(0rjzXiu(ZTl-H1 zklC5Wcq&;I7WLv@#0v2Xn!J^`-w1?gNowiS_pF}|%aP5(Q(6}o`y@mEOhgn6-RW%E z`L^HP@EHE;2B=+Tp*|NqX0Ux^9b6db8l?kqKP49j(7fMOU(U<)s&c{OKboTx)rim> zq)+<6i#l4z64C_3nmhTv6TlwX64G4*GEI$hh^CYKqN_RjI1d{Xx|uja(=MdgWbn=C zt)M&EnzM4VYls3w-FYxyFAEMzo=qVm5sWDaqYR_R(&w9B7c7#?uf13=d7V@T z5X8z*@o40mDxTzVj)sVEKsZ$B0%^^`V*arB_E; z5xFr;PEUoJX*v#6C>>b$8ls;jS?&+FNBFtiu=^zjWU&V{zfu%Ik})0u8SHF?xb^-U z_dkcwhW*mC4*T`{aJ#Axq-q3ZR)m=<-tjhdLYOja_|p~JbQ`G?b9-6uaz%~l+> zo+kGW;yiccg0c$xvdfaa>0Z3eIsEpeQ}A?k)lUjm>fEiS{eCOtk#v7la^vv< zsDc!7bfwel^6~Td zNY5CW0ur=ZUVCWX``5iSWLym`06>@y!K?kHYM*1j5WH0J$#M#$#El9I`#ltacf9LJ zTzU!)jGU;-d>0dLILoTQn2|LHnCdD&@~!%CfmjYBq~>3Y0(UvES*_qQdc}j~^Hvw3ir&jeS#62>mKfW{P&j9)Zh0&LQWYSFy}_FW)S>+F z+slM0L(cx3R2`8C0#6^j@JK5Sd0K>!Z84p<|e2;<3sj=eGbN<6or z{J6~K?M+~Y_5Z?t7iOWOObMm4tdum@44~c zXhb0g{>dXq1kEI_IqYjB+$QcGY$ZW^sA@?haR^pmvNM@F`)Z**p1tq?ec$K8$=<@G z$ph&(4Hi=0011C36zg!&@6jb@2joi&X0r8W+Vh_MzF9$6qU=!$#o}_wm;<(zmGFB( z4Q3uCFIr&81NTsU=zA>o)~_*AN20^amOA>7q~W}E0C$udp6d!x$O@u zQ20nW1*Ma*fDJyf=1dYy2}d(1fjFSQ_-h8sMPr{KdyaiF*vNYi~g5)iF;k~G#!ZUD$3Ssv3W z94I??7O{w(SJl&6Cv&t$lW)7(v<92WseW+r!dp_-w2@1yMq{Bi?NaT0PUG4 zlF#Se|DN~NkdhEGk(Y*Y>-t9Ncp0P@{fdlqlEK#2(L4+Q%QMrZ}MXYVW5>sqmB`$S}J45t!1xrUW_iR1Kq?vu zz9@5VF*((0I^H`b?lGy(uN#dhXu$wTO;0sIq(@J0D6$sjW#9;H|H8r{XL(&(&o0_j+er}AJO4Yl?0fj$ z5e$Xj6R{eGc##A#>)Qv(cFG+mrMY5u%hkfShUf($0#~r@Kq;~Pi>=M zOYiqW7mTwcX)Nz2@ieHSitKNqUH?VUeKL8zzr*sNP9Tz`t*qB|=ks4R84o^%5TyW! z{k+Zd#uexLAFp!rJZ`_vhHkwBE(adf9U^QB=MO~ z+aUs#t0WcKlbx%F!=?==n>d5f9dL`rT%@JpK>!1p?bR_n7pllgmv&{)7Q8n*wcRI zL6Na`nvp{}Bui_?j@gZeuthpqz@Ng*d|9mf=G zuPpp?rkG~{gb3uIE;`br`Gl$-+FYKrd$KfQjqJL-8*StqWAd;azewkrEb{@gIr55T zx2zfTB_uP$k&gGeZ8#FA8Li4HlWa-_#s^lP3Hoci-$3RuNE3%1u(%MqkKz05A#h!u zXBZrgFq-z(`+%g|8y{&e^2U14^IlR4>j0LC=qzHTfU%0d`R%2(bGw?JjNEvGhtGoM zT)zRZojWV)saf^xkFLz1Q@X!d1CqRvLBn0MXg_Nz`RH63Q>j-?wvgD=YeEuf##y(!# znDs6}_cEn&yUtSfw_h3owl2T|08T;J6F2(^q&ty4X&MZIc@IrLT0crKO254k*+pZ5 zP+jE26amd1ZtTJ+7tD~X0!r??q-y+p5d4EhTn^}QIIjB@P<3pfgffBK3x?Wym@bb9 z(L?1>Pc&Y)KiR?lpcy!}*B#>=~*>)Ziq>tFGViOljq9v{8+31<)h7U#3vrzYNKHk_csNmYn0{|r zPCylyj1Bn*Wv%z!eWK>`nf|Td*Xf(C3h9ZwJW1oaOLT2-Mnq3*sIDQQ7C5q}@9`xR z+Be&P6=P)s;RS60ew^#;c_q2_o53a+Q(bHNj~-fS)Ra!otn{Vwl-Mtfit<&KSF3hV2{ z62whYMlxy#XG8NAo5E^qtv6Di6OGodV*q!Hrm)}xiezN*OBX4+tUe|m872l)l$Q!e z*hqoL8f4eZfX~r1*zHep-p77?XwatT#HTr75jk=3BzZZ`JP(tDaX?8_frvl(^Km{A zJvfW`g}jG#2!F{8Rs!|-y_@53Liht614!v-2;4@#wfJnl2BP5b)3TEJfmp#X)vEp> zgL%IVkm)dRl*B#x8qN^>MSshmAVlKZJr#V|QF|4B>7y-i5s(z@V{oAHB6j{fW1n*i zuHlp19SYp_YbFBiTm?h0A@siWJawa*^t2d7%2l+R63ohn@>OE9D(TTPCinZG!9;EE zxOV$fADfzPoY%gE8~`zYKtV(FR;;S(mLCQOX`%tc*-+nkm`dTBfTLsGp3@!HMTOO3 z)%U-z#gHLJel*vmEW{Z{bWlhk+;LN2Yif;ghIm$^>!|>Mi48PW*)qw}nwXsXml+%R zHBAE9r_IZLioq|`06uCg^q_)Z%rqpYFvyD=O0-+|A6g|C_cOX;g}gxG-6q$)XFL!@ z7okoE1&y~6&ijtAg-ZR>%Q2`o;l27YzjX2DV<1cE0`KT?mx4&<-j|e6zv!=YGsG|{ zoh?}TDDJa8N0d7}$;1*<2AeMD%i7O;V^UUntuk~W2!@FH#+CTJAl-hX@7O87TyYXUnwIPQ{yDmY56ed*&7Ej8Tf#Q;zjZ>K1q2$-|o&&TC&o*Ep z5i#Kk_@avicv#tGNvaxJADFL(gaDX|NLF(sZdfsaQ51()ocj~FA z_vH^!B;Jj&Mk2K$`(Xw^!n*$GczQ^^4u(JezQi|QLCCd1r%aXsdRLx20plK+o6F*+ z^M8B&7v?i!%=H@B;q8VSjEB)ST?qxX9NB$*lsz!zo>*-3*jatYyXg9WAI1WVa6B z+cgRQbL$@LV>=$hFI|1WlwQ-Kdgg~vRxPLY6@FAJK|$RZVlf|zs(KNc z@BXY2C{7xjj1kNMypXuNwHJZPKgc}`fA;}o>rIlK>msgq{Qzc=`ww0pu?R#T5Hzt~ zlWWnRubvihMW9}bAE8|Fb57u{sVl&}$^)ENSi;1KJ;R!IAcaRiB2^mMSA%ZD!3Qst zb;W1?eqA8^?3@^(K#&NDJm}xs1a&Tlk=;V>=?aof&o)X40I>Wz$CZqI_*;Cgj_>Jw zgVZbYE6OlPe-DJrq7Or)xeuD|Dst)L zQLpj#*6OfFcDD-XROdbUqYc~ci|a7{;FUugYl0RJor-i7y+OD~8_(g5L%x$CcT${j zmvo`wK!@cI2onL>sc39ShkgGwl55;D_>?$u>){H!QwI9_uk}Jf6QEyb2t#JaZn^+J zJ)+^Cx3h3gqm)I(20MJjFS>ZR|BUFyE|XYr?kik;?T~>)QorSuEVx zrzR2*;=EkF4Q`xroSsVImU|(wo>E=Sn1BXmq^Q{Lc>}S_#_aKMyd-aIKJw5I8;hFu!qIRF+18l4YW5wM- zY=KRDCNXf|LCQDeltZ7k_i2#e{L4P^*$OrI89}Fn{u^iHpIql@b^XRqABpQQk1KS! zGlEgRY(Sb{XFS90dmSIszA<2IZZt?&!jab3&Xe5@zx(<5z5@2;URT>G5P{}rM@TSd zIvOrhqs~f99rW;xx2^be9gv4_?>iE!i8*uFU4C-_u+5JNwGX=i1A3va&lhzW)!`I6>z+_*@o+$u7q+j+Xf^!jGUqJaV z6!!uC=GiSfg5e4D{Uz!hB$|UGwj_lsDgO11*h}fxBH^>VTAN1ghmJ93-iUPxXus5! z#fKA3|JfDjcO>^-G3T~NdyvjIHQsOJiqi(l7J?-@sL+=t)M6G)nAv29ls6Gd;^pp> zryxG8Bu?vT3CFk)P^5`@z)$3t6r^ZNlHVR7Gb2n0MDjp1?ccskRA{S$iRRo+t~o zg!aA+c`vv#aoq^qJ(h%fPBlg_knyBxm%felXg~Z}%d4;qf2PB;B67Gpm#uE=5lOf7 z{B6(l-bqbD3|*Js`(skaGi8tg_}$t8@l}4*hF5@SuKr91_mc0T51Zc5&9h-X67CsJ zM`30m*Q1^MR<*|_*FTR-3a}#S`37{U%*8lphi;&g^Icei9Os7q5fC|MY98|n;ouV6 z4Kwd5{U*U(@u!vCZ~ZqEDu<%`!t{&mF_HL*qcVAud)dDTtO=?JmVm3f$j=K_xD}rP zRxXah-1fJN<7>B>p82=GLJL)J=mr2l#iHXFo|40o90u`DoW$FG16q5E+VkH3B8%*f z>eqUOsbtY>iljh93{w>1`#`rky;Fwn&MuGJMEbdJc;G&wf#QVQb655E$3FSZ z)s63GTl?iXL?zKV><80v)^Lz>fmq|SUvN|fIX8Qh5bN~<84CXfzwWDBlfMz04V+t% z9yqou_$54LtQ%ik^qysBnh&}dWxylw$IucvkAa~YNR2xssRN$5^>1y3!xH-G=s4ISZ5jOV{*vSQaDBk)gjh4 z!%GVrK_{SW^@`T9kVxOlV@#3)hm?ma@41ETh~Ag>>UW?*)4uNI0KK<&&fpQ4oIoFP zOY|4KpIGZ&T4-){8ylxL-$z~gKMeO)$=pig0 zhR`zGM_q~=-364~d@ak9AC~g_Y+bmx3Njm0<|1oP_Vl0^4E@;1-~Gl{OW59*OY-<| z7fqgS-r=C!|DvNB{euY{29d^mA4#J^kfVX)hKe_Mp+;;`yF4S_s72QrKkxB~L=QF5BdT@w|0r_wOjCv{#&`1P-(#}eHx0iL*y)U$}-?dN_`(sIb zfO8iPX6eu0j|dm9IDw!Cn!Km*^fOi2qyElj=L{#z3L4x6)AdG#L2`v@ zDaMk*x%wDWt3U6-BM3!sw!MCe>5;c1yz~2O`JLkCf(sE}#I2ndvLl4(0lN;0 zg-Z#Dye3N^TR7_T(tfUe$iyRHUE_C!C0U6bcdR!NswbvMak@Q`vN=WTV3=KiwT>z6a{ z&g-%td3r*ySAAi6-IH(bvz#$1ltQu=4KBcyj8 zRd#WV`1yq3C~vW7WG@MA3D}?3z+aie@ZQOSO;GL$A? z87uJZ675fqbg_2)=+N5%2c|tu!WB6?W@qg5lA_XUL67Lwn<3)gK~hGTe7C75%&T~ zu4rN?0!wq`z^j&g%F7-z`2%XC>J%EWG@Bq5SQ!XGxK_nDo=N6hotw$n4>>el=Jwet z3ZX)8>&|&w7j+&)NT*dy7I;O(fqPBoup^VxM16kT6<&t8Hl4&%3y60YmXr^89MmSv zI00Sbgy7`jFW5g=fFWpN1mYh*RIi4H|94H|IbyPV6EQ4sh85!gTrOi=B-jsF242_k z$y?yz@$6SMa!1mipXEtX%c?Jp3X{JKB-+4Tz9Dl3g_BK=uTovIuk0u&8fgA_dJqzNdpOJ6x6Mz;y)z8S>3Af;yb`dHJ* z^{U!mwX{{-@O7sA<&y@pP?`JQAr*=K%6YDG1e+&pq`zg{NhGG-4$g^NV%{pW-!QK8 zwnscdCDIn|CcCTkzSZAFKBehw|Hc{$x4@qzzT}Gl-@ZK~{7JONbIJyLz?7IAitSoDzAl?o#bfDZw< zxp=7+pszv%M;?;+EzbOee6mJ;$DHAy3h%eN`u4c7_T)rvvfTHd12uJgv^{G;y!00=vj6a`uM}!cLo@~I z>uI->dJ}nzijY{jRY&IyO^%wxpzM@=W^?a?o|wV`hP{03t6C?}_+O7^KE5()IQgoH z_I}>F@UG|;p-u8-G?2p)VatG1VhoP{NP~2v-AbfMJ!3ZhmNy+@zX@z9M7tZL_r)H= z&2<+>7dt6IYVfa|&8#$YNZ(8dW;2#B?#YpurlgXJz&J||yD2Sy`%lx9)Lj8i6>>vW z()10CW^~p84#(Hf0X-aVG0>98YOV6sVP7W-q}bnU2ZC~t}iX+VGKFH}BiyrcW9bSqT1x&N9M)ZzCh;)6_A4Rq;QxV*bDY%7N(`Ejd+2xO)R- zqJu>F8KG2c;^8uMOBf}EIzji@qxMQRkz0btyx{TF(QQOp(mT?d^$1bEhv#14R&A;t zrYBq{b*||{H}vEyr6sO(=+jDFJTtAK6q5GGO7BYPC8TATSNY)pp?*!}hGH()D+Bxx zIxU=nLs$FL*eJkWM&b$3E9Rawg@v2h?d6P~tHBx-Zg+SJidpq}iyi3_Uh$OcS9^!} zb_n+rAyOH&Zi`~iYlBAq<9ltWCB^?Nu9wgBapw@p&wYJ~eJeH-m3A24jRQ~68b~&Q z&8INr=A3Zf#y?6k^MD)hlle>JG_3ZFB2h&YGJ-#lXZW`6-MO$h2q**Qde%ilx8dKT zS_*~C`AZ*}=j!jLe`9Tv{L0@zq4!%Isn}PTdGG|L4?a{~d2g#<%k6$3t#3`H?Q_r0&lPY^ z(rwZb0lNVI!_)g5%l+bwEF3I?Di!m?!K|?YxGgD0blu$Rdl-$U3rP*r`OTf)GRn?k z+z0(Nd^oO68bt%*r!-+rVsVO%G9T(yYe$`8D_pZEaX_XwR7NtopJuPC0xBpfXXKQ_@agmwLn zL3o&-Lt?!SYhE+X`aSJM|5+}q?6GfLz4Li1Ui5@CcB^6PSkq7Bt%wDYA8)zQN6@E+0 zL=-rKHC*v>T2E{;rO|?PE0G>pD17p<&%NnC=H8^)Rd&nP`!yM1H}V9MKt+?f22mFs z2SSLt6cV6c@4-4>-s>CT>)UX{&e-{#y|tEVc$x2B)mC=%}J|p8^f!#1~D$K46*D7epwr&w>$v5b9BiLTI_Qgk^(|ZJAzSK z5B_z)xCU+#m`i*NM~6rTDFPpZm}Wt}5^ z-0BDs3RP4<9|UQXz~CUy+Vn=?(-NQ{@#vFybh0Jy!Iiq>x<)sS-&_wU3UN_~5FY|8I(M_)K8SebhgGe$Yn$ayVA>Ybk$H)jq=5Uz|Fz zMX?~=p63g-hL5ZW8Y20Gt>>JhMAmh!+`~xvv>m2kp8(WQmV?ScWrhN`Zgz4}SukYUOdv4x_8{S$tjk4r;Ro;Wa9*_@J` zfF>0e1;I3UH+cf)bJxlhK={uP`zNc-6~)X61u*lU%t5TS;6qcKF>c*cU*-XA{@&K( z1wrB9!hR!f$}w)>mBR+`Do+mb50F|z%T#_iAHS8RdA?+TO(#5`cqv)SA-MHxd9nEm z+;C^5kf8W#MvLnTdB~gde%y#ToTJl#&zN(Gh9_gh0kbl)6ihi%2(9k?ladXu-%14q z93LwZ(t0{U-#l)-Eo^Pz9&${CdrAk$hhnM&;jg zOn6d{m(JMA>i}sxIVd~c_R%%P8x_THY6y$I4~U86NDE`|raLXY-^5=@;vv#zh&e<( z{H&gDw{PXzuwx6NvOux-}2_cgYki+QVhp{`$}J6)KZEo*CJU^^gRG zjm-d4(1OieevsIoAl%*z(OW%gfli&}PfJ`l=X&6wA3WqU?sZ06Id=}w*&psXlyxPZI{92eH7x=w#T`j~ut`in6$N~9`jbe~M&-KSff(Tf}k;JYPIOf6#t{s|3q>;7L z{0J6yA`825Z4@W7RzE%GcyeWt8vPTwcx=!GHR9X8Yr~DoS+9}GK z{v4e@KehL*_2EERSHCiP+-hS+V+iie$ z@2({+QaZyTbpuHGQ3{X22La`W3u?n?1+J#n(%AXlru(Dbzcd63jJ%6nul1_NntF9m z&x#E1pon-F&;kv}x_S6`&~}wv3>OVkPZzvUQ!mHji}POV@xW@n2z~-tAG}GF&_@+zDZm5Au>v zzV*jAdUD;4o)mJd*6>Vfk>c!EBAbk5P3;qre7T`LzBOr}0OrwwFzLuOU&?gAg9(`j z#XQ^))U4J1ao-VN>ES0<7BO5X&-RYP<9O0U3P0f9tb<$^C?0T7BPjw9j9QG5Nmq%; zGK*K+vEFo?YkRG-aI?ksAM#s%&E=qCKzrr^N-+)^w3+NDj*iMeJfi4&#xGySN2$uF zA1E!)!qdK5CH!zFb27IE>tee(EWXdeIt_SogGylgB_awM#F{s;mrI>6!2%lhkHUNc ztY1xkpNISh+}TTG%@%&&b9on$o#T`OuTfw6tS_(ZsKj{svhLOqLhg=)QnWU^IYRYi z32Si7xfIdu%2kDGAuK}L)t?>P_Jn;U0<}5b<$#9W;qZN%uC^ASbFhfI?e`CnryBl? zSdx7D0eMb4whaj~UP?<&LH(A>&Br=I=Q;@T;9jRKzBnqHT(1(z+d6wvQTKxFe=z=9 zK)LupIYVz4cwGzrVn%>No^J$_rH;M!f&iVZ=SrqR;T(@WN(b4zzW14|=0WZIc%b!^ z|0-d+S;ckTZX_{E#*b`aka?wblu77eD6f6<^t*d3{Q004#4_@{=+jR@i=nG74;Cv9 zbC2Hz<<&DJF+OlwC%Nd(^jobkEp8ys(5nv(!Tb{c-p<~O?ykRbh{5+FJJxS7q5!G! zP97=pJGAv7GPN0M)Zp}mItx**L7hOm54Uvz(TXY z^8cy)h`3@YuwDK#m@&2@xj(+BV`^&6B$vJOLDZGq@cjI~P5ZHWMtis}^qxSpAKpoM z;VP`cj0dEr;gaYGz%3bYi+kdSvy*G(wl%i;5siHeMfbmpavrlEQj?`%Vi zFmzZ}&e+cj{d`SR_8q@Vb;ZiVEnP)q!v0wbg0=8UsiRA#JGAk)&8`>YuJ2cZVn1qs z?~z~ntB^%Q%BY_QuqR)C#m?pNvGL3X)#_{wGF+~3aioTqEjVHQ{@j0+U$@liqFE#8 z)MJ@AWwiFDGF%q8T)`SBp`8))ke+ey+~Ef#sOQQ3Tj@(?B83Bq?J8teNWUq2R1dWC zgD>*{fk5If*m5z-w|uV@qC;)Zuy{`eHldjV!%c?sQPfxCcjXseLtrsHA0j}W?2QJ` za9{S6htCNec@t0FNBP~k*qsMQJJ#qiFd@L&^m#8aF!zU;1veS7X*s+FYCGlwjX9QF zgst5PDImI(hdg!4$!)E-Gs2W?PQ@750`=wirOIYm0vsfaU}Aja!9fiW03LV&+2HfI zf56TNh2&4AEEu#{&8$E*1{ zTzT)R;>cHmAS>C&BM^*&fWdy!A*k=2s|`4{X z?L*>7NskLOS+EapO{rjF*QG=? zku)>v5nMnSRlFWctZhBb`aP`7)JXgDMIPS2Bd0H0XrByKh5r;T@vsfYA-7dZ_)yPS zJ6`Xv`Y5S|O9vYOTr_wfujzX>A&=;pZN8yDpTs?{(_TlR19iXP^P>%T%5(vq?OAPj zx&?KoqdM0epD+d+eZCeRTgX`ljneg&`tcZ4fWr2=Fo?18SpZ@$WlHwCdT+h(D zjVMTfh=^C{q6lG%hkp~p;%e!QK2@urVQhq(vCyce*UL&WQlmYWD^=YG> zdd{zB;5M~1Dn=FV4|m%4XDkx^{?4`z`^-|m`c;Z=F%<={hJYiOlqlP%CpAFKkk+1?>=5F)nkU;0$kt5rC)S7Wh6RYp5`{i>M)~~9w z%jCoD>)ZZjIP9DwqJt`$*E2;_9azC!#k%NR(uh$KES&%W{Jr!cW~ts$s(mS9KSxcp zgd3E4M|zp&!~RIHjh*DOMTDz9%WM493Bd+XawU^=4&DHrsRT@wXekW9IHVbtw!lwd z+vb2W0ZYx1vJbj_k8U`I&Eb0I9gf-5BA+}e6yY~W+O9F9|2ts_1j@=Skkh&n;2 z;}Ztf^XFz^KpaBXxHO+5OW9AabYip!97Ip40_PO;o^OMToj1Prt>dPk8G)e+5^h;2 z3b7;!k#~g5LkRBril9Tjv<1%WKaOgKu$){u#FvrlaB@9F`P?JrmJK6GONO<=)2b=Mi_aVxUZADh@bzoyllZL6 zH6%pCFjCkI(D-3EZ?k>JTP}q8-G_AeL)QmG=+}r z{#rgVYErWS&j%RW_frNp2Kf|k%{z^hAaUgMOGo6&^Re?jd+a9)1Pp!PVwW{|owCCl6UEOk7wjECPq*a5$`?x0f{39RdocgxpRG{<^3ya)uYc^VtGU74W9U za<4iB;4aA;O^t*!*09jahJPtSUPKByxukiw>&Xzr#XwpNXE!k}uhu;}Jg_G_E^1j?r492l zP$v*btB_rMd-bNB#wVMB7VQqnMa3k+wBsguqfwO<-FSbzT)O|8u2Hm>5CSOPo1YKf z^;1C4(28-;7`N|Be&Nj##5LWs*QE+;Kloa)=+ShjfoSl%eQ#J01mxKR6!hb2?wEoa zw)5bkVU)D}fXKMFn7TqypttGt@2!9vGHaWA=-Hn?jJurH|nz zIt|}4NORbOW(dsHp-B$9)x;Y!K?#C+qp(}tzTe16kf6iI4_9I~9J1=c9m_si+b_YVlJ;IvZW6dujPq zDecmyPQ2_oQFY|oVfOC-j25WRRGV3R20Qlr>}r^aoYcW}AUqp~1i&c$p*HZPzS z3fN-bz6cG={N<#n%K<*29zMP6@HBnE+HahN?HOZoSZ}>n9%X&=YpOU&wEJv?;8IrL z4X!wQ(QgSwbeex+iI~OP8dw8&!Xk;j^?SX}p~#PqPk9fZrAQR9L8OCMkVDGKp+$!v zWQS7F9=U5}tiBc1j_y@vXMPl(d$1F*glwn#dl|28XHq}58A%!D1A(-y40WHsI82_rfqNCRAkBZ2jLQDjq1ZTE|VN2!u zfX)C@I)%oUw>Kaw&a~I+> z`|$|6mu&a7HhbH9KgMu#`(ZbFntrbyqcjXYqBKYfrBo7`*Eya^s-#2IW_v{Ne5lt- zVe4-@!6eoHnFk3|?M@iXyES$0`wE$61p{eZFI1yAuo~=V<|0{FbXy>o#0cqLo}X4H zzl-PWgrBE5TwN%evKmnD`W`KX~Eu%NuS-aq9z(jL_5yK~@EUggWGQ zm@gs;T^78@!~4|iIHn9@{uaxVhs}2I;{s9ZfMTouE z*dIfE_39^;%wY8t-#l+pRUE*O&?UL2Z!h~RegK`oYhv6F=o%Wqc5)1n>>mOtQ_&5& zD}BtZz4p-3ubk4gT=;A~0PkxfJ^dq6m_gju@xVPEpmRc)Av16`%dnFyPp>@g?6;u+(qnrKf{;7*C{y+HApc)GH(HsXIlrKXj;KKtFgnbH#|gV=j( zu$Fy3M+F;I$KgweS*9?BbT|9lbP3={eN5%=EYn-J*gAMN4fcc74T3Sp+l8l>Z`Q+` z-aW)e_sU;n;foOUDLlCDBN73;8u>ird7eX8>XYGU)P2u>)`_}&DGzOz_OT~Rv!(TF zgyl3{D3Z|b4VvW)g+oy&-tP1EH+oVi$xQV>sYRp4P~2GkI1Y@pH`1*0NdjoErrqc- zPcA+~*AVw#o&`pC{N10pjP9Ui=KM^({?27W-`x+XX%})eJU80Hf_C7q|DdHDfyZXvLJ>U^Z~Cf%xjMQ~q5rAG$eZ%91f(j4bdC`$U}aps(MI%L`jN zcGq9fx{&JCf_zH$p|g5p$a+M;Tf#2*5b*n{ByEv%v7oj~0qpug&Wm0F?%1%wQC5!A zxc2*w9G-^?@#~1*H0py4DTnw40B*8$a6#gAFscav43t6* z=|ZEkU^iifyh%m~wrLA`HTInixlg=|NG`Q-Npu2cXU~)yx4fVw=bb^Y@tmTXhqrpK zDPiXI-^zPI;edd2dg$)tf4)^vdHlOGOLkUsP}v#<>}mEWUY2th!2(e#wf4l{MirM==eMW>VJpc zc9A2yw;tdcIw*mM4!Vqp6H6?F8CWk-3ufEH9+Bd@@AK&%}%EQ|TDSkZ(IZqo~%J{ z;*1T(RX+MhWLzPz@U9;w=Yriqg2G{_3LW?!S}>ZI^%j#2*~^E4k*8`I(q#ar&P_y* z3a{H8YWkpqiPc30iNu38<#jw0<*M@ILMoJ~=(M7}qc=-EzfxjZ5UcK1d!4hM4IH{G z&lkk8cbbOIXBVIIT%kiM0Mm!yX@%3pBfQA??M6MV)DOI3a$>CRhI&$UEvn{^d)TH> zdj!10V3gzv(uh^$6);rTAAl$bCF51O0YN;4&3(DwTNSu>TsBx#-IvdL%cb*;Sv$@^ z^m!3x)Iq}-6y8bRL|T$cCi(|000;Ix6o80Bq=@BFfFxQXa48?$!`9-<{2lkZU}97p zb?Eor7WMd;;3OtroQ2HZ^D`+CM3J2SG_rxZ*Nr+53BNk1aO6c%m641IV@p@G`f60muEDlj|tGlHQWgcd!knlj-)Q=NCN!s-gQ9tYT% zx7DBb)6NDECFfuEl#p^C22LPkKmRsnn0pnrNNak_?mBzHVxBi?*_w9?w9oS3LTAHMz*ZmhNT7o zaw7`v;_E5srx~Ky8gxL}ACYNB4V}K*sf{ z=`p5>nO?c%qq%o8zEtnv(YnB0X0O{fqg9vaMPcvWRs4M~4YeV^I%v<_a7mt+vA<7Y z`~{U$edJd?SXO86rpliZgqA+p8LcdfF-~M%*Y!r9^GdJSlZTSKQ6fJuGpj7vFN6$RHOTFVvVH+=6^B6;-LAniGkK| z%S*qbQQ|X_<1(S22eH4{w-@;&*X8z#r)saf3QR25Q9_l^e9+?@OPa51wi!8bzmEJh z1OVTXTT@XkPJlb$56zPf3B7!L(%9vF{MI+O?#JW$6`k`%y>pNR(r42xmL8Jxgu8-X zgqSnaglJ{IH~j+^Z}U+7N2SH;Ls1tl3#f<402{^@TGipHF=J*s>WNF?Q=d&n>r2bo zBxGom(-SEw1kJAdVU6AmG>Nk*P=CnuL_67ea zBPcsjcn?Q|%1N*Ii~c-v^KxQ*2%G=4u>R327(W&7cu< zL=3k2Y=+OyfXs!q{yIneS6A-Q+AwkSyyp2cF-`17hcz@+<}na8sBn2s!&u596Z*AY z_j(Sl@RG*==t+Llli+?;u5M9eKJuw7M=-8ya%08Uy2xl={s^=qw|C5{^++oygK$6C zqrCpx$LwoB*9jSaFG)~@#XhX)EXc{p`{kdlm~N@F4mDbSXhmKQx9a!{c*;CGGT(bM z$n@Ok1+=iGk1FqIm&eu3u{dD-NezqZ=%Nu(zEC_CV!pi zz0d{m42Iw0{LRbm}s0d_}z@lL`K%(zS zKUG1)QQAM%g+-k1iY%}OTOI-&DptyPo8gMvZ@`bm9~~b%?Ht1XlLj~Xv(Gth=8Y#}ASU4uV+hoq9j= zw>edeuy#lC=&z8|DCU_xc<}55Ea*AkDOjH<3F?}MYk(yy>Ya;z(An0ZGy;5+Wt*D5z za|or*B1COOk-qr0Q?SHj?*|X!Jtx>Cjc|8>cf9Q$Vqm@GDR;)t9?mILNaMnGwPfPo z$Q}_;P@|E5xGTl8!4BCX#WWF7sAB@a0BE?W%6lq})@jsaPro+Kr}=H<3(9xqzZ>fFWedo>%bXC%?agUB+Y=m=9d%*ZL#-)YMhUhmA_pZD-u zt9@tZV0rmV0uQOz*!$^H2IVS~kZ_GEYB!{s;ofd`BJwjLH z7<~YuVxf$c7t5JnnSu|SAwU(r!m285z)FCHm(s=Y4{s(S^g^lY(;zwPY&?)k?~)KC8mguG&{yvCTrPj-c-M<>;x-=_%Q&CctAl^uM=Vo<=Q zg>xil7WFvZ=p>V>O1(aiq>z6gCY>!O#fowZlasdGR5T3~n*-L*0O;Yn*WV%z3_(+2 z8X%Bpa94*{y-2bjPuE0`(#MP_SGm{TB*cc{bIf5G$DM>(kA~Gm&Xsl8S1p7ce4Dh} zMP?FXG6p#R4tn22~SzhY%e-uja~ zRS@6z`{C8=5I5wZaTDK=%jW4`m6OUdg@2sqbE%m;{)9r&`vNp6C>QqO&==a)vgrEM zX%c~eEH2SA+JwzW-;4k~j%TH}oPld1Hp&R}^yr%VT#HR?%mQkYZ?x&Rgxx3`I0{xQ zVzUpD@I-Mx@o-Lh%f<84rmKrG0ws*SCD!qKW0t@HinR~HM7kje=r&xq&b~X%8iKK~ z(uz7f^u3FY^*J`F@y-@6VgWM8WviRp7r68Jtavzct$OJ9YIeJeoh$fWKk_PHBB&6u zP~6@1lIBuU2i;J0WK{;urcYS&1oju~OWzNa;0FpW58^u{#vNZ10MJp{$sJNnNwiUY z5B(i;9_QOA9*}R`#1xEN5JE>>>mdg1QbOXwhU(E4+}3Oh?X^6`Tj&W9;(bPVbi}I3 zsWa*ej4q*Na8DPtk^DQD$r;P#?~BlFt^`k{;0TT35$Nk&U)dzT75zvc-Gl#x0^5S8 zGv2I6K2L|)A_-)Em2m@x08Edf?He!lfSw&EwfsS92Qe!xPC!G-Zb0J&oFuk678vmf+g`^lsN+=v@Sq`S@3MClE0z~51v~kY z50LC8f21on%;+8Nn~#O{;~AkMSXz^7lb=U8l<3! zjL61B3+*mhh0a({&yn2*ppO1>C-LS#Z2BTu)Wjx%7l`k2SvNteT67ul@1RNlig9A$ z&--7XI&~paTcXH4>L$#eiWmOe}oy8!x$ebJ2OR z>}o4$7Zz#Q3H&bdmND`eflb~FMhrqo46fLd!qs=EZHdam;Pup&95STAY^y26sZ-BS zaaylqj$hwI^6YOEP7GfM8Hj%HQcX)z5ui!!+1x4oA++GUX>@Tr-X6R6xQ&w+tRfb+ z1Cs{<&=4i42BP_gk{fjS!#m1l=ZEa21CpUM;j2gc9$!hG(!M}u?aaR(P$fEhUo#$s zT!1q$JC7gEN?jGCAb06mq_fz;47aY8k9!KIvn194dR{zx239QgC{oC}tyl6iimTDH_J_X|- z8-Vb9fsO<=pJAe8fv>;XROF^OaProk?D*rv_@C|pZur+_z&&Pt8?%E%p)sDiT2lu6 z=R;7&4X!SKM_n_g_Wc#}g#zgakGC6Dd8RtUb5LWJEs+HRSq)zHu)ZeCB>PS0K8WEq z>yNGsdkRJidxqZ_N`bRgF15=3-mC z#Pr98JU&9EgCCD`aswAbAvhgJmh4RUoL#%cjK}gP<)=7F2=6p|I0pOo?;DAX)J}v@ z!T~tGrROIwLmgFSqij9%teg4!`b=^H6wJA%TA|ukmaJ7O!SMi5f&Ns9fBQwCW9p}J zwgc?m@}6RaQMx8Xf>mo?&P?6L8CVIn zM^%EFUhoc|G!VSUd#RU@*DpS`3leG7w8a?sqEiPU5puE=9+B%lFSM4xD;Tv24-*o7t9GEydz@P;hhv&HgL}-N+ zo?JskPTR}4U6d=|P-pIV2i|arbXzka+DJ#{;8?d{PiEw>MnHvKZC?E?U&UHq&y=da zZ-~nYacR_fgukcPjyS0wO>TQAr<0!JB}g`Y-|`v*n5{<5vQfB!cSbz`GMZQg^xb(Ud#wU zHJ1&KBi+sVcA)in?2M?oUO2imR;~R0b(iE6dotXK|AlAA+-f(C#>=Pw`V%eWI*mCg z>{4!-L~K4n*X73~wEa#`L+N$!5)Ny?HDj(9x@$tV0@J>ZH*Jk|)yAHvM07$TL-f_J z=j3p18+eMrt7oovHGka-O_lc}7f)`ykQR&?%5OA`JN-Z-&RdiXCySsIOFDro=AXEX zON!%(Ua8A<-x2cTM^4jntZe!1NGv9Jzxq=@x!e4H#HyS4iH?2>2!i%{HtYEpVNK)W z@BJ_y9;~TbQTGgKD?e(cMNaO5{Qx)HYa?94yaDP{klE02*1eHn{D>=2rH?|X=miOu zqLeA}y_nCW08Izj9(V^}9|Eq39sx{TWK%YD|9TzQR)oJB{8zx6C5fbrg+#2SCa_<9 z$Z<8IG(0+O7*&GUL}oeP5dv!NEwYfJ_Owv3v;)4KuPUHTQ!Gp&l z3-jkCc&GOUk!~6aPVd(CCe(DGVj_v=!kwT9XI5_|hQcC`p#`*Gayy^RMC_-cKHLQr zI;7`a{`#)4To9aGCArvOGBzL37Zb~cib458>MRS8M;%{JDg8lwbU;iG=WNyYa#J!4 z$_p6@-hf@h$qJd$ewBo8v=}Mm(DA<3^x%AyzO_L4z&Mu^1Ki(m+b8@R^Z@Ys2yogU z8Z!i12xJ0*lQtU&a(oc_bLgzR0AUwXN*Sj{9T42}ad>0p?()k`a80gzu_m+uI(v+U1hv4%E+LvXd<&DYfD=MCWFx(%S4~Ji>l_0=3iFdo0 z?dN$&Z|8uyGQX9T&!1&OENU+4@cx~H&*{@#okJC)Ovl+=musV$9aH2fE-5G6o^26|duAFOU7YkkmJYuD-%h^%?${&k;AP za7xo>frSEQEJwD~hu$|1R1VdkBIA{bwbID<^F!4?QSGO@b(UlvZOAGQJgcL{;73+GAzVC-I70hr1=R!n1ePL!Ggz-YTXKIs1JVE zktvn5!3%lq6~Yv}fJ%KjQJByl+qS0ZmK?6t=N3!rOM)^@$Chq@uFGoBzfp>>onK&x zBERWf;mHHLJv+_=IYE8C(K^bme6#1vL$phJiu6P8NEVK(YzRkV|K zx}779O5ifc;>ppR3Sq+xI#y~uj1vM7zIm2`vPX<4fdFXnu}83f=2Lw=zaH5J{U-qd zaz@TLJ_@J|7D6j!kMzjBqq;{uE$8dZZhZ3q4_gf%EM`UL9M3KiP*jCC_wv@A5d&P= z^dEx(-@@?qURxF|8X~zEDFYc}O4@I3-;OEAkx&MVsvrlxelDWm-t1a!2FY^Yrz}{{ zfbSD1mFbHbLxoHK_<@PBwg@#)RO?j4szut+RlXs-K|&!g&<@K4Q+> z4|xp?qF0kfmR3#{=c>P=dW7_ma4Kh&+h^hxEKCYdnn?;lgfbY_C1Xd>$A0)>*Q=?I z(h}b5uha49uy*3d&sO7Pm<3ooJcLh_#v?}s`EIVco4rTYdrJ-%&pERl2J2yzj~ns| zzwbCOP@gI=8oRK#Z?$206cL}%J3?Ja11M@u1(_Ruy75 zeaPSSvyU8+jBMPzzTBQKomjCVqsz>~+2twleEInsm(DJz>?AA6U^h!y67acAg7|b z$v9KXBnHMC6;M5b^)!>Wmm877uLCF$Ay}i)GM9xre0~Sx1V%o-5@t7A_`EKF0OGN# zuvyuw0wx*jD{CrfM2ezScT}^d^giyv-tsZ^=vGRmSJ@8cfizn@2~z|?>2AKA59nDzgl*rd-Z}TKnuRai|a0glf3q34tnZa$KHfP z9mW;D7hF}XQ&(s~MBuH#&fwNW$rg?`dXJH-gfb?hh<1ojCl6fo+AKFU)a1;Vpe?yNlImNfx_qN?FSDM6~)OKMTnSWX7<9@VMVi? zk*3YLL!o|32gZXb{5af<#5AlUDhWn<6;DZ0f@j;;%P24nSadZyJjf?L?T0q!E|@`M zV}_3C)*)Bp-E@1)8Hiwd!yF1?>RX|P`}u`j0;L2@no!guB;vR5s_t*fF~O>(Nk`VH zJ1ec~!wH>_YErBb;9?tNqix;7{&=T(TL6685qMrWOX5HZ>yX&Tm0gyPao3ic(haegM~x5n*-u$lq(4xj$1x>0s2_N%c!W zEBk&09j9fOiz&{rhwW%owhE{*P23Z{vuS33JT7OB1w*uOK0s_Ih6OuqF);-bowhjm zoGVnICIe==W-{gLO`;nPS0|pmKT-JaBs{kl$G;&_X$S$Cs30x@?GDz>QZx#i-*wYS zAMD;}mldt#mzMq@o?Unx^!zquM51cdx5A^fO$>=kFoQX~n>c+*4Q z#o8+t|L{`BnxwB%=6ayF-d=MD3B#8eG*cmmYAyRhZD@Fj%yk3)b~>N)Ok8?mHEuQN z%G`~z86(MG2vl0z!mW`)nriX!aC!))b_R|Gj()tV6n?O8N#yXQOH5nq@cl#6E={(@ zg(t&0`ESYG_}_9?qcoEmBcGLi5~_WVfJ|SzDrG8QdLK3Sz7-ph9j{d{p2xu=FtAzY zfFmZ3AXI7As>O!62ctdv`Pj0iw7r7}6+U8X8x3ZN;8R+L^PGbf*i8gXbZKk7^H75^ zDuMDF6En$*OND`STF~orhKGU*9eq0TZ9|P3{Nm4qWR3?(sOV(t6D8;2u_%!$QbCeB z+=&MlqKy0r(1AgiO_c`M%1Kj7wMR=2|L1O8c1He+{q}C0{-893H*qHmRnz0Q?C+6jRe?(00Lzrw?U&SP-!Z2GheXAA7^RJyYdwC{ImAPj!E!llE43_%B;IYfPu^Udw4ip7K8fL`>d zQ5&R|sFv0c$x4HXl_?hjh4B;N^Y8s|{Wep6zUOB~Uu!Ygdwb{%>kwo?0K{||aHl%4 zz|NVt_idBmXFG^81SilG(D>#TNGOCDb&61I3xr>EkK7%cB9Qpe$DUk1pwjiQQcm|p zsq*@aMFXULn+W#gfhp9nuh-B;*_1LG(jZrsuLD!DaW0wFi)awN)|y z?$LWS`q_v z;(u3L4~-T~F3)%TeG`1~jT6vCXA3o8kQH#f=yq)GmrI&e4g0ao5Jnhp%n`2|7K{j5 z2z|YP*tngx^L~YS4v%fwgE1II`%~*HEJ&XSBC@ES1(s`7tVi>HRBhch$43%FkmT}= zPOnfrQ-@MICLnp@aU%W*V)m~RQ?Gm$nCy}PODFHPqfRF-e4#2bFqqdwxVLaY(ij`1 zGM==XzPW%6Csicn8%}d|be_+v#0t;cfc(rB{kQ|3(=Z?xe&wmfe^r=sMMnlFK;}~Y z0F?p`2S#KLVTP^UTfUeYdMCzJf8O5_$&+n_E#5K!{m6cYZ&6gEA~>J{&UR^Rzq>1$ zj!3|A$p=yT6ABYDqeE;17yZ7A8HZM0D+_TssloI2XnyhKw3x(>P`cy9OKFNFH-AKp zd)z?rKK$@?nKC#_}Y>em$4cOxoaUDhU!YavZRKvw+;NmFeLM&yidV?Ao zW6hzhmA9o~AX-%o0~Zl~3Nzl#{R-O+b3y!E9OLlIu0r*8A1G62$twek1(?JU81UX* zH~igPa{VD5Mh>8dQP?BJo2P3g@~i)pnOlRYZ`}0Th>g(ItbB8R-qLhNp<%N=LVC94 zg`~<+*RkP5U|^(;?~u#)H=};N>)R(4;B}Gmkgh3n{>F#`nNQdPC*MArf+M1VQ{55;N5vIi7_P|*?;tkqxw z!DgU_zkbqgNB^f6aw@zsZGybwAO!>l?$_qfz6fYumvYe1AfEM-PXBVAohmHC!n7Ah z>MqxY!;TH(?IPJX+|*4QqCLi5jK0@C>y+;gedQ7+Tf*kC%@qX+eR?ZX-6_hr2(Yl<%Il zD3l-9hTyIc>A^03*fX-L{RFq`6scY1m_O$&do+vGUuFYEygP#`HaZ!>B|$4h^~4tR z7zAvfFhBTFr0F~1%0Lpz0YW(S(xJWQWPhHe4^k1)l>ex=esEy4+2}#LFkqh|w_qFr zGn3b}{M=6OiOBx;1V`4-=QZiyb{_Et=6~7~Z>wO6!*SRQdnijEyo?sMtQ2;L!NB_l zK;DizK25Mil8hF5Asyi>{HsQ~&nhaX_W@i7HV=N>{*is1X1`2h36AWB5et>|M+Y8Z z2hN>deS8zRJkT}28fq>V!R(BtQ#0N5c@w{Hhvp06l12Pot6d)_jJVJ^7KCENNa)2g zECT6e^+DOzkw{+`Zs8sgX0wW6hY@}a5`VYR`Y&uGU$w^^o3Z2T$g_ocIT6C~x}>4^ z!wG;K!9y)9g3eKSi^ESaLQ}`7BpgSmgdKkd4THGdR0BL5p!j*3FnJfQHoPso+32fP z)zhx_wju6wb00R+R&nqmUGCXE!U4A@ooD*~9o)DU4uX9J#~=DHHkmfb*O;N5>8_^6 zVmYLzzEP+Aikku=ojjJCTF|Hlcj3FUB4=Hfq2T-m*|`F}h9 z>x=^vD>Ok-|2pHM`S;h$zYf?4?0-Fw|2m!U*Z&=NewI^~&Sl@>-tYdiKMGzxTBgJo zLC4vOoAFP4gZoWGGGT-L@0(6z-rZ_q5V(whnzNCr|D68w@Apppe-IdMVe$H(|2ZMf zO3eB;o2L!G_s`Ru_5kDkiMD^9_HgeP4mC}&3}IFNR-ciKbe*3 zU;jVVnYljVbbNQ?{}Jl{xzm{U=l&)Q;;jBY)_>pdp10_ZC-{n8w z48V(X-{PzJuV0{>-TZLk=6=Af#oXdm;ygPs?;;uY2W!`SN`~TJkKLaCul_NA-r4^# zPlyrB_a|eX`$tafhs22iaAlDGxqm+-X*utoKSk_6Dw~Gwn|{I#|8WHC1HqH&{xtK{ z!)o#0fATy(&eVB-;QyNgYJUR%pC_35Ir-;Z;c)r?@L)ThuK(@e`sa8f|Jx1!m!s|W z&-eB}9BtQsop6f%|Lg$6Vg0{4!2av-DBC384}Tx{`#NFn+6t^1+%&r9H*$(;Ph36rS(6=4rE zWVlf5XWwlvzz|h|#O+wIbDE(oo_1(Prs@4!ia$9c0ii2czJ2$qJ`aQ__gjBUkInZ1 z`hr&Ri_9`k97boE<2UWcAGKdi;5Hk{S??sYDJ1LWD3af@0FL;v=5 zm62ZflBQ*wwxC7YcOX2?EHr+O+$ALD{15-p!kgxCJ+@~I+M|WTEIY!)j?X{|A;XRO$79wztCJGLW zmIyJ@dk$m$R(WcUyXQo_AfYC+x{#a zPFVnIDa^mm_kT|XDiO|sg}>OJGE#d9Siyw9$A^uSD*_RuqJPic!E5_T9$4=@@0wK* zA)vzb@LZJD0pEp3FW~G3ucN=eqysW=@s4m@3ALt!tqPb+9tm+!fNYqyAlfNPC5A+@ z#j}y*kRvlHvm0HB(q+FwGk9!9je{1I5p$s!b>+J(Vl57{b4p@N-7bUhI$Y-diop8b&MX}N586U?;h2*4Q#F4D zKw(52+6Sqd9ML(pAa)H%U))l>7E)=k@xpF8{UU=Af9qC67fg_js5K0AS(q}zkd(i( zSPUiu#dsy(YO@eBVsM z+(R!S9Kre{?pYEr}AL_AL zmLZ=pJB)uFo^0TNtB4i`xIY<>kk)R(a}N$^ccykwG{B-=L>q3x{R#5}V(t)uMy!}H z$>Glg8qm9Nm4Yc8<_NBiG+Wca#=^6kBOFL(+@)2T+{8}l!NyxFcU<~E-G01crG4Z* z=$HF*@K+p~8#DBaNBN2of=H6e=>T;M25o4iJnz;fT0Qyg;2{KDfGbunKS1^BMAxVT zrvwWmKdBcRFX{R( zkGB85d%&>pGYzJuZ%=Sq?PTb?&x6;7aEe_ zt2rvx@Zm*uzN_drB|vVuIg~TkoVT*wg+gDag)%XfvMsL%H9>#Au8};h$U|1c3 z@XAzZbC}|vU*lGoUfxg44`e9_Z+fp;Fb~)^D*bSi;NVMqM{KfP-|$K}9wessusI|L zBLx(S%Rpz>({fcxIGBRyy{2bUs&)c2H}cVEzBNb-aRc%LQ}f$sw!*$$-w*UTEJ?sr zO95@{O(Cpte?wT7vC9o9oy!jfQK8~auf6K#agy&=ypR0tN!(y+*3vuIQOeiqs~Kkr zK-VhLm3~NfJGAQ`9emgxa`46TJ)UXw%io1m zHw~rVpau1MZZ%aS=-|R;XV-uY3E8GmjnNMKxS{)qJJEl~Hk5$MFA}KVxsYPNK5UWl zo!9#S5BeH9bO>iGSI-|5vfn!S3Lf$8?Q~yEVa}4OFm1{pT+;W zzo1v(#?Jua9Q5791~NU~9V({Pt?p1tFCel|R4r+!-YA&Y(FgUVs0e=ej4gZ+xKo-z zAdj`njQyxP4XW>-Qef?u0(j+Bb`%c$CDmW(2l*1-tlav6Ez5Qwj-*^E8AUL+^VUXs z7hfe#2}0|pWpEu2xHhvI&fzzI?l?}F_d_uhF}GMCs_-S6P8WMR|A`sAQHPc^1p)&E zkMJ)qy2gHThY`eioTB;hJ0n6i20uwd(wqc5%Hk(rLpQhl!p9T#3Y*U zbNndXqPQPlV8Pvm%;6IRKSRhX@<>fu?h`oqT8uH<-IMHJ(lnm};)PrLcniApD^gQ? z0Pz@%=^*#f{#E@{FV(bR>8}Sqlm*zUKeLpZd^Ur9g`DD0u}loEKKB8AmClQ&SQeoF z*oS|7ot>birkRSORZ52K@4Q5uLue*(aAJe-c_)FB-bcq$QY zALMrgpZh1F&(x|~S~OYKC2WKnrrai{Pt>sk;L_Qm=S6pjHc*6w$}Sm&H2^|6QxWP`5WHo{%!xy!l3z3N^jJURXp0=L$G_GQ9U1$E3{%Ykr*`#tEA;$d^@zVTj< zP!e!daxfsAD~4u^urBf47iFT{Z8l6?x!xKgrlxa_isfg$?;&WI8eWH9(?E)?kA*!Y z$4>=yRmC>niRigVz3vr%k$hl~;!YlVo#U(@k^l1B_a8y88swrL7+-v0<$$>4y(6&j zH*gv87O4aA8u>;K?sVqK+{8fy7PE&7x9cwa?$0ZzzrVM=Wrwk`v}Gqf$j%ekT?ecH z*a_1L@*@#k_C}a*hmP^EK{v@njDP@;8hM9i|iRfGoKSj?4pd>iMeEM6MC7E9Uin(0UdybT4#W)FtcpUbU?3 zN2xztQlhruZA(~rZaNI{QH1Rjyl~*vqmy*VMKQ=?GXAo_4%flZ@(J?ci%qN})8$Dx+T36JX$T#gx$N;W@xm)Lp zTT$+-j8ZX!Kh@Ozm_Sk;rc8qL04*{=k;mKQy}1zDh?BYO=vHvW5QY!SU|M1jKvl~h z`n+RJ9htg z<(OKcK1J#GAzK2yvNfY(rsjTsYMCrfJ*BIU+4sgNn{WlU^x9 z$6Pv+n63ht%4LZy6;of>bx>}@Sk+U|QOVky6+g$1R0@VubjnKuPt+x%5gG}nPC>SQEgUxbvpGr+MGg zZm8!j03qv2=p%p|9P(vw#z?Jam&U`;A?Zd1pn3TJH2{S_Aq?m{x(Bf7M7O@#k4?)u z=M5(5&_qk_Cn_3#*DpwnK;*fvxKFw7bUMNOU5*>lCb9p&1f9V_ z0ZwmfZ5eeSthItoIne;|wukyYSkUeu_0@B8x`!taC)Xu+>8|rW@c-5eLF%D+7kf2JhGE(WnXs3DCs@fP5_`k znZ}dNXW&G-U&r0OFhbQW-wsDk#oD7cNC5p-me?pA1<(}ZHlTG;P$6k**(>J|Dy8c*_ zMu3o-0k_k5S|5K|QwS^y0eYDS^z zV1;zR)PKrzHvD9#1G7mpy}iNm0nyG@e#!fg;mv|y($uDl@IGSfr}ujP!CW0%a)h|c z8qq3i=*qJ9YJ!BW5haY@kJ)A>xt@lE_7>ZEa%}u%VGG_}JD*vBsmuD+zs?lI_5ra?jNB=rg|IT1>*OGQ3s)d~a!j#hQt@T}@WP|$f1kQL%>+N6stA7x z)YC2yJw7Bodh;MPIReY1%z67Q_*;80<7)rb;M5+aVfLXe@yDSaNVg7QRnaf;0fTuo zoWATC3fNFCaXK$oG#-@0;e)nZE7Z@5DU)EW90!n@OB;#FQz2nQUJc>c)F7I96|%Mk zc6B()*ta&NCCN+NRMErK%Tq&X^N%NU`H8htqT|$xQ4!xm>#!@1P)?(<>iVc-_^8Y3 zpLKOfzkI=W-B|^#L`RhxM*QF_E6-82rJ)| z0Wxyw$&iqFuaLI7#oM~c!Ha&=1tK>SaDcw2=j@M5#K))S2j5QR5a4}!sBYg+ z@X$~})M!Rk9n^hR@!$|WbEUok{<4HU@FzB=q1Sd?S1TZ90wxDf*7lFxiGQEQuP?PN4cD1jP@*%lHI#Qa>7@S4R{Lvbbn-rKe@YH8L7I&Qpr7=AjkW?b*P4FOn_ zL~hNyE-Wcm*c;R>kNlI1JB6awpiv$j5Ovds`39buyU-xX`9Rkt^pek++lvTUK zC}cs{$0MrLgmVrz?`sL5Qg`Li4k+A%i}f=#^;h?NgR6)gQ;sboOxW;pOA0ocA%QtC zDRJ_V&+RDgr~3&F6ojyiD@+0Lbrcf&7N-bNjn=p+NR5s^7`+7KX(It&Z8=RAyfqb> zqazgUaf%M>hC>hFy7tC4^u}A_8 zxxWYBG0l-IWZj%}wB#h9GQ^M2G0yn~2{PL4e%QDFs8NV}3C^w#TTE6x$hmx2NaT z@YAM36bR=+r@~O54$|zFeQFDh%+*aV!s5UjQ;-`8^#SJ$}ThBL?6SjqQFO>{J$p?qd(g_5= zP1oTMkbb}~wN>;4IQrS(0V$@0NFG~8&*=!vLKu5|J~4B}OM)cRw}!}H|fLPo6Y{ynh2AZY!1#1iSjpL<+4y2+Ct z0+JOY@MiG)6zciKU#LLF3cQMNjw9Z?}eFS!z&b;FIH!{HQobkG(woDIU~zFfBf ziCfZM2x(vUL>6MV!Q%9c0Bc>riM(!Mm{Wh1+;2dRLme-_x9pciBE1py<_lfWy;%xu z8>`K~=p}>+35HrGSm4J#S=5k0QIw}h`M99qL?r95JW4mcH`NnH5A5}Hu1&dEU?J+- z-QAI$%hz7`{J0*@ie=hV36@mv~Kx; zl}+Zs4%DmHNOFi`KVGL#N7$`f?u40Tg zbh%xx&$+ujz~%DKj6p0;_wM4_;d@pj2bc{l47{RFs7jDlUe$2Doe$K4y24AqQn?~B zcVElM>vlh;HSyA!TdZ`13bDdRaWe2ngih6IU5y*k#d+nD1! z`k2KX@fbXd6`rSZ)$H7R^{=vcQd&;Z$fNpR-CtY<-)7T29GB44BXQ`UrR%Jtx8x&4 zg5-w3&HZsK1SDwnA`EQ>kZ_=2$7R)`yZke~?c?O=(#wC9)&~!F_k-cv4j)KT9o0Da zklvaA(qO&e>RNT0h5no1RBh42Wgpv*&y9*cK&97?dH4CHE|Mup&mJnN^!*`R)eNf= z!B?lbJxE+DT$^@#95JVLxbM)-IDNdjF4}j%$Aq;W&!99$4rzArwZZ^!9}vZ4%dV%% zdZ?-mrgXqJql@6~eRW4x?B$YnJIpSe1C626|39xl1B>!fQkmmCaNwCmvr?AmC zWr^KR=YL@;h0Ol(ZMjUtc!dAM6JJpF+U&?W&l^dr^3y<6st%1rL&%yV`n~i`c4s3g3#9!6$YBEcIG)9i@Cc4c72tfSwgNlxtwe67yZ>Dh_=342Ls-wb}e?H?nW{tt|nMSR>6 z7d=n8Q*Qpch#Y?DUYEh07&2lI|9l3JG3|red!_W3!`mc1HSXHeUeP&t0<$XI;MHis zo%&6<4J11J+Zs`p>{YcF8ZmrPR&<@2yLvrj?T9LHuwp z6$G?+M1&|{6_tEV0(mq=7DcB<-b+YRbZBjcxkIV5s+bOy4M zRAh!+)e{17waRQ4a&EMFXaLbfHOF76`0Pj!?Q6LEY$nzS{~fhYroYVh?9@zP+cM%g zhZc^5@G*g7aAH&YSNq%*+J@h09}iN?k-;VUCNAl_Ve_E`1U=$db{~^ z+g-21;C$$}8Aw=lc$meJU>=?n3CT|o<{Zlcr#xq)R(%>Z3D!z7eIDW`v!a}mnYo1p zMl$^ID9*b2W6U(m3KdTD(rvOrP}{{~5VG-;s6(6_zj8$&<5BH;dy+6-t&ogiYb_0> z&jP(k88jg7LRMe$s&=`^FzE7%GAeUd;_zEcNX)$Y?7)C z#7rsl)w=NGlMP8RXyA20c0$peo+L>joAx=`;}Jbo%>20rr)=*B)Jy|c&3_H`X;Wb@ zcpXfOK#Eps=OHp{-0rv32W~(gNKh4^qu~c%rcz$O^JFks93GC89BJm3uu6xjd_Okf z+t-uy`J5h=tsro4pqoH9In2_TZ~4W(`HkOIf&0r1=LD6~y4|2#L33YlqlK|mMdAz^ zlBLe>(C05+CVej+uX9}gw0)n?T)PJQ@B}{|!qA!$#34<`x!bWH5dEcbSNPvd=yVwr zI_@@#xAuxKrV$C%%f|0Y2bqiH;R;_+%Lp|U#IaXu_bn*U0Go0i31hCRS3&y?xaSEm zT-Cq|4NLfZ{qBHtOHStY5Z}20*{vxmpYAF1$X})%3E*{%ata0+)q(^tVo+i2>F;3C zZIoJF_iu+FAu29_-v>EW1&36!!#ug&_ZW)@j`9iDA%(Pnx%C{nfW2RZ$O8l@dwxJ5 zZ(yz(yVU6-1nL>@InK%3`PjUG=L(}5fLt`?>L&S=Lyl=rYg`k}7t=0KIfV1>dyc$y zN%(%Dkd={=Ed&q{$jx&dljOW>ROE`GmqtbhM+ieB09oTjOk_g{|46_tUFNtoM_p-+e2K4UWSwV%ysTQLJcp)To@) zEcp81DDd0M2}7XKdk;R-DDG?Le!}J{+i)HSt3-5eE22GFjSZ1kM`LuV$ zpJIp<-VWTsZ`21QG_V!hOZ?A+%geJOAMS9s zT|i8(GcJi{8as(h0uB~q7OrnTdOUB$e&IFEB_MApx*wPB1ohb6z3wZHalV{H%l^H1 zWi)RkF>%>`rhxsnQ5v{s*tN3s?3e?a8MDq2ePdk%Px5M(sBZSpwA6aR0JDNRtTc^~ zCLKWsW&|*g;O!q-eL?9251U<|O`p~ml{Jq05eTIQ<_AmAKsC}RQt4aA(}TtSEN987 zbbAZjM>MbT3_ICPNq9gee6Gavdwqb!9K6fM{-WtYJPu+dYW-&{h_o1!cSO!CRF9|c zc5iZOs3L^6UeD{4(+`SG-4aZCC}hMm9-mEv&4JZ^u-oOIvS<2>mKCdJsVlkzK330D zfrHfDWJ3u0B*tNxZWOJskUz%%M!$hPN!npYsV9l`_kJQse8$w z7$2Xv{Q~_+p7H1+7m;ID*xSZ2JYaa$@?);55qgeSr)EvqjxW!mrD-Hcl%hIc&`ddk zpq;Rf^N5-91fsufy}qpjN*EQ_nX=~5y=3Pykd^OfIVKY(ffp@B5Gx^ z>!;QiM1Yx_fz0M}b{(9f^}(U8Jv+h$VYqk*MmJY<`S=@arbYt9TDw(jft~|c2AWPEkxYuEQj8ocsZaR=M!m#h;%_KjM=@G za&{&e9I1S{8{ap*8VoHIUU#Q!QYk&`RT(>a&|F4S&&h$BsaM=|dL)qq2b)%0oa7zu zSUWjyQpfp}o8r(f3KGPYg5K1;Wr$VD5WZ0>S8N8ib#duDu}UgMUNo+QEwdYZ=%jxBX4I&FFbwYIu{N;{@;~{CywOcL?_9x$hGA z$M85%N4^))3YtQJOb+z83WnXcr0SfigMM{I`%!W_KoA#D-sbjp^*={~><(*yY}@;! z0psqWp@};yy)_T<-|Ug4B1oS%aHZoyc%e6MN=V+UHOJ%+5 zH=>~@HX5EPMa<&?9ihFOl6is>@BZj7CG$vQ?=Gi1YCfJ~e`%orNdWsB{s@W9&ah&r z9JCLE3+|bnc~Ht(jYwKhPT)fuHc5X6EuVB@-mol)f48J-el@?p`0npgKR`6?6$!#< z;u&M}R(My!IR-U|Fv2TM7;_;_h7^3=@Uuy8NZMXTw}BB1s1pJR^Ger$jQfSSpJ$BWPtE01Rt#F?e+Zc-grVw#`u-| zg?jxi2y>x zIl4g20Jt2~{rPPFN^DRxFWxxL{>&YMe05*->bUQkasN>F4~>L~|F>5QMOA6&5;Ydb zrRF8=ftoa=9=ZiSI585YW_Y9ctumo3v&sN)(xs)M?bD;P>%69FMd=?mtJ zXA8bC`n&<{ajc*k%5eojc>js{7(exSOzKY@l89JywPz1XXz!DXYI7x-17Z3neRxUQ z!O8$Is|JkRV@xA~$tP7g&qh!8(Bcu;DnYd-|+McVYl-jr7JkNBc(idR4G(>nqN*yHGV9u316eRSbm|B6VNDlXq1 z6S%!WT_IC6$Gnz#sm^mi$t^M0^VHbSE=8I-N^hm;(qLac)CqFT%hlcM1Eo$)iz6)- zCn7Mcf5Pjrmsj{3sMYWH!ugC483 z^*J9CWQ2@r*lF}#|A26X%;1%M`GnlGzXIbezn++JubCdohgt~$&}6f)ZhM>kn*&UW zoJo+Gyzdvmo&1cIPcFfn^#}_qf|J!@iwvps2x7;*YR&7M9dTU5@3L1zdfs#~zz;Um zr=lC{P~B-Dh0!htd|t{ecwXU!`gf#l!;|X<2-!sYyly* z?R1b%I5?$IS7JsB_PP0@i$=$xcg0FFgYUWPCai_ZE8cR&=Ccndm!%Lo8qg#bQ1sbA zux5f9&Iy~1EBj&(eP^kIhRIZ?12lq^_FxjvRg-O#%D^crnIBmBC;6@s`IV9`Zdzm0 z%I-pA@V<%I>|tM6?MRb8miVK+1y<)%cVutuzU$}{pQD*_|H%kFIDoc-{Kx(|phcxN zN7QHU$%}^Qd)-UV$a*$|GpN=#y=VVw(YxFqExcuW?Q4Vz4BG9g17@)FZ06<cKSt>TF3FCet$Rk z)P$MaB_1lc0TyaAU*#Z|+nb!1Tr#3}dc@i*?I-E;oV;<}vax(|Z>Hz1TZhw5%W z?;wDBAwIKE4H))`&z2>ai1xf;>bO*$-Zp#foZO>VKknJ`5FOhFm3`&DM(uqe5m+v) zxoHsT6^rJj2d*doqPJE%@ro39>#!Hz6VISc0aeH4?0LezA=cMtBjh5Y5GBELYbfml z?f{3_FBu?W3?7LLjPm@3V?UXWN^9|hzp@P0TJ{6n*%9>y55@W~{G^BUnWHhxYh80g za&v`{kmB34SNm&!&K9nW05Oi~jHe;*&6)}ENQ0X*?}#pja@d(BZgUNr%Na}4>3AFf zIR_>Yq17n`xy$qVD&-B)I!dN(vxg#7*b0pM3Ly3f(SROS;tF zF<`Hc7`a?l)ZZ=hD(U8n+Rt)0s{3RhllYQld3Nov&36bxya21zlpVfO6Dn5+S=T_; z1R;BRQ9)E3UGL=*u0Lq~k}c7`Zts`Pbcwwm{E#&oG_R$~HiIo3I}gIuF+ zX-Hm>dh^5#r~Flu1!=Ep{fzo`V$P=+j5&}Q9ViZ%FrnrS@CjM3vCwnBGdsije<2Lh zd9PyQfayH406$tyzMvjj;m7q-$VOISk(25(N04^b;6xm%CMZ941HYYwY*Vg==znFY z*1CSHOFD61XY1y@+hNPo3u62Effkl_3qiOz@fUT=*T+UK`(@(tS zcDDCz|1^~e_EH{uLJM(#dh2e7xYIoPA}=}w>ncc z+$B^lm*Cf<`ypTPKFFY5`ADcPM0efxG^6cV@B}ZOi0gccQ@_5T5$!_uG2nyb*=M9h z4cYDuJGjcGzjix{5ZLE>+4IjG$q~$nzzyI2=qc|Z*B1jo?pGdTq8F5OW;=D|A>NdU z2DfK%!8OEB)=utLpSrfz;U~qRW*n8j#`lQ)_3yqCF@!aGJzOP-omQ`qq9ejWz%EEK zb6hU(qSxPz{)G_I&{_egW+s7qsSj8iS&L}>1Msi&_PtQay|kZ^oo>4P%Mi}(9haf9 zMR?4{(5qhinrM=X#z7cg5%$4-meIvqI;P+6DLs!Wf z!?{fe}|t%9aLp6Qv3$^#}o38z-`Y1vk*O`9LTcV05y5T_JJgVg@@jN6Y%l!fwJf6Rfo@rr{#hjVai&nW-^Sa2 z$w@v;a{x@&-0vNyd8!>!gh1gBEkY(y6Tk5qOmixTok))V|EMq@b?0k;G*&A)Ce9_n zidKN=9@G@_Yq6hz3X#Lp0lPSUz}VN$)Tc=lZ}Pp6kbR<_9{5#Ev6c;*$+BRt*M~^HCLsO+$V_c< zV6U0!)+ARK&nM!T*)i$)lJK10O}^PRr1TL?O!)9{FVi)fZe=-|wMSs0PuUA@R9UMG zbt0Qi-JkXa_%!64u+9pfzx51_Giy(1m9&VtA(hMk&G`5&*#aC_hgRk$%W3%(zc{^m zGMJY?zcc98-)P<6x%!0g4AgdbRbLg$yP$uNJqY>wt5W;NxQirE)ZU8|jg!QS#4ra5 zUcGd$1aDK0#K3&Sw*#cOCu`zm!$xH7ufiY7n&crD#xVln44F!{2aUKr)cta0NL6be z+21OJaVU3*N7upThOe9X1G0Hx_vk5pLTk}W%)K$r`V;~r1OU^w|6({GlhFbu*+a<;zn38H>)u{~CFY8_|OH#`Hv z{_$nA-?eSo>O4hl62Nx2ZQ0+JR$H~&bTM~U4U)5%z56gfzh^+t49EGtp^5$G(ci?s zDZa%NNI;T^77b1^3C9l)kME_9NU@K(ZafI_5dQM-j1D93o9ts;G)A`MS85AV$!!9yu9 zTbf#KZuV~vC(a!*&mQK^+d(9TY--GK~ma& zK9#fmy}!t1RgX&m=Q_#nIpq9veuV=DMWAny>+s!&POmN+j>UfEjTR26bkvH2M*)ss zl1qoQTP`j>zw;IE9{*6waOVhfcJqOvwZGn|HokatzF*{&b+9`S(MsbGq6tWX{0(C<>3_UA&0kVcjxl(#D}}$qglLVy@Yj)03AaaDMCgLzku#2d9 zSqYp`bS=W%AvroI7r1RMm3(x`oc0b7m_GPMM7(*RR|}*y0Vms@p7dSz>72moRT%-Z z9-*57+e2%I$1?G-o;c7CF{%3%IrQPPfU~R=KB>%zChN|B=vw1i0eiY#Yk;sHR98S) z;QL(k28F0*T2Tl1p55ywPBuGBNgoz{j6Vs69*EOC$zbQ5Hm(p_$Q}+0)*LTABEBYU zuAkB={fy94rFgEe+^mSPa}yb0ei%FU4W5A;?et)fUXV>lge3wv@!)ZcmO~Vy_Cca7 zyb~`xa?q?fbS#)Eb}t|vK0wiAKHEYkAbhtI#tYSrHw7Df{m}~RFrFU8&V%DWa1$`H z0a^J_>JB>cU1&2)_sEI!VnuSJAp1vhkP+Txp*}(*t|*6zw$}%gt$}whfjvYBUt5aY z%g6Va=b4ghVq_&jw=4mNu8Q6gp}Wh)QgdK?jJU9;@U55Q<^83~>Cn*o z;eEQ?Tje7E7AMpUh><6D|FVAIID24o!(395a{5?B8Nvc`u{VHy`2^J_w0-qWDy~O2 z828yC^#azL!=>KG9F_F@5yQ4tke8DY4q)i)5;!tABSy>Ii`$hw*J1y5jiPEmx+L}~ zrFrhY%WivNs%fQI^a06OumNF255D6ISVwJwf<&Nwhne=X^KFA>ps3&+3q@P7p7{)1 zEtDVz*TW7i@5h>9UlWUE?`50GDXU;dbV%<&eO`^Mk2na&fxJe`J^6zB%htxh@r^2QW99nO9CK?QA7wKWv2gqphx`oa3 z_=KI&Gz(iQ<}v#tdMrX}_0O}!*!o`IfoTIfdxf5WQiP`m`pTMKY$5%TewVwqFzX?~ zHLx1yZ?t*Q`66GpmFQ5QW@@emj$Y^29o%3^r2*~6{e@ZH{m!^*2o5v<*W+`dfiK2i zqxJW2t2{S88&Y&%g63=yYD_d z0bHqL14fk_FVa#N%CPw)a3UUeFuna@$A5Mm#;F9-?Xm z>I%XSaM2f|n7#RSRxepxfnOX2)>Db%y)d!&_oIH8=3NK_hJ*bYa4h{pBLi9%Taw#q zD(4T;r+rIbB^va-%W=QlDFE^yMl;csQspmW>d}Hf>l-i&kFl1K*DT9DL!BKXDTr*8 ziRBY%ghhq``}SX&eXxX zXYPL*8Jy`b#RdD-6g-l5DLJDIQSJnDewI}sbKZ>(&JTzQU55~L#$27Y-sUNz*oM8o z51cN&fq!aAO4V+(xOByq5DyIYfH3b2`Qo^@a0qFWv%W?m9ep>}HiRCsUoq*>;}RVo z?f+2g(AV@_#4!kB`#|Jd{NkU^%PjTN`+fcMJDgSHCJOnOvsU+u79o4<0Ww(QvkRX2 zC8=`Qi__0?c8@W2<D~F&R1WHt2wZa-dnW!L_8zTHf;7Zo-#A6x8pf@LQ1L4 zC3278<3|z3hh!pC1KR(xa*^8XqA790sef5J_}h8PccBBX*z%$!fT?}DKB?XXMjV3j zLLvu!@A2ec4{u+MxNF(CUZnuW+kQPoN5@L59~>G&jRD!_zwr7L)vW9^#GK>O#Oubt zN<2JIx#ush0zU>jE&fdwX| zB!2FkM{b>;*F-KQu_*{VM5sCi!tS(xb&NcX6#a+)gQe>|UHij<*;nvIN`>rnebKpi z`RW`dKdtBDcG`kP6I-YJJHU{2LxZ8!V_};%>~?Z~u&?9YBw}2L<^Zo7{}_Lrzpwo0 z%-Ywpakf6U%Kr$H>K?28#n6BbS+FnBX9i!UbADXL6CS2AlG_6{&(7xfowN0nsq2RP z4F)Zd=FledjEQWI=j1&dOi}QI`x1NY@@g-lMFb_U`sL7TH^%a(b0_@-1S=JwArfM| zd>}kP`e~^=YFLUt`$Bbb=s~YnZnsLP?VrM3%R@HBk|`6w7U35r13`cco)f>)@}ZpS zk}bErh)`6;_S*7e72J&_2llq&6tmo_n-C-00HzM`8xCO92k)m6hu;~iYtp-1YV4vT8viCz$Mrp7wEtVNlm%wkn;R5iuWjyP(3c#F>k7Ov1 z{v`+me+d;u+K3)sV$i*O2Iu6Dr)mNdean)eS;F`}>RKK65xp|9Gu z_s1J;Qrc@aJQW#nO;mEnH=)F%tUxK`g7NsODVY5OdKU-0--C<_4X0;J--GkDd!Wp@ znC!>mgBwEvzNa#M7$Bv!AkVJ9Pd1(5u0C;c`CQJ-2`l8MiHdirP5X6^q26_!XP6R{ zdC?uXB|)P2u>F_(=C?O+90ykatLNrxdQ9F^5{TvzebN21zlqfSVn=i@GtUFODCkA7 zXn#u>&#zCWp~<~}Cc~5UE-kp=adBLcYv5F$=O(^e@$u)vc}dO%4WPHljxK*h&%?fS zwT);uCRpu77wR9-l%lznF-BETQaubbbN^I+x5 zVXu#ckLbN-|BRtStm+*!-qsYcD*dAUV|w(Z5a~S?(q0UG39z?@?B-GJLAbq-f5yc{ z#3?%AVAS{Enm4f~VNs%gbMU3}{fMm%{jrEwkUJ4U{%tNmoMS>cnhC8UzZNHkvwc;X zwkn_oOZJu5!Jzt#jCjq%(yQx)7goiWTICYJ(k&#OGnQ@s6~9vi{mOX$1fovNV|BHh zxJmhVl(`on=f;8N5u>ydg!#ZbW1bNNZrCH+rOiGQwcUTh!enZLbXx}gqe-J}L)bAyK5GeDW!ts!m{Q#n(FfPp*nWqU#-uV-r)Zg{; z3y#$6Vvd045D#^_=fBpewHu)1pYE-^!8}0~xS0gpu0BkBa05T#CI>|-&h=Zr5`TqZ zB13`BAI3ZQmKlM`)iv1LW;xNTX>cTH`LzoDrLiCE{{FuZUD{Aph(Q$dWh@a#Mws$~ zjsFR7ZS!vt8d)GKBJz+ml;Gntmer;mAU!X*gu61&8yv&JMAyioqi0H=So9`h&;Y0E$G+%%w5j9?+m->#G5033i7nZu%;m(3b3 zB;xUECqldp1uOkSfELUA{qQ^zq0>bAyKkT{S0W<)dmtPe39+qtlB@M`xbmacN_C?> zk3ygBSO3&p!lL7TXi3RXtwq3i4@!$|xc)7cw^^qFiyDu?{O^jfYNXC}>4-JY_h>qJ z*B(tq#9fz{d7_EY%z{0uM(8o;>mHIQ`%Ok%9h2oK95PZ3jtD5Gd4&8tZcDB@i+Qrs z`T%C%m(Qi&0>S-{xG)bYA984!;2G%^+XC2-`-RI9||;h zfZ40>Pu2XPVI}{S5!@ngz%G@uiz*5fg@8iq2xZYgeF$$KV#;Lpsyu{j$cBcrX8paN zV=3Puki7Upk|@7j^zpElVAI}#lV9OptSq8Nm@>k2PGqZIFjyx?r|Yh4x!mxR-HQ)= zI}lFlj$+5X3BFKF`U%dxdKm5Sxrf|1y(;dxF}yROM5gxN!{GyYNXdwH^|iu|LnTDr zKEPta97P2|3*_TO{b9}vuN+v1F3$y!Ekhl8^@aXd2}h8`5zA?y{sRC;p>T4XtPi3! z_WhxehKYVa7?^SloRr-|6&9XPSbQ%vv@qU|PReI^`_h};$R3To%8uB~h>x~hQWt2R zuMAusNys5hITs7);v3Z);si_4BkN~!89>n2P^>UBa8xFF<>{7f-mB{8G=}#~rhfo{ zNx%QGbY5GEs!qUtX_`s-?bB#@+$15;AEDsrUZ(i(g@n`n1fUR?K?e z4dV0ec%f576|Y76AVo7O&Ji2dwg1aDjrI#FhHs9!h+*(ru%(?4IEn`V)9K3g?af2z z#W354IG&$RHne#stV}&${cp?6*Ehe?-L~cTTfdH17`*VC56w}pC zO=)h=7K=9(D$;%3M`90diL?Z93l}*3B$<50Mk2&;>;*?d_<`W))`cM=Zo7uxKjS6V@dg2s5nE9Xm126Ie+&Vf zVOCI8#FuD=%xnx$t--$C&GtnC*2F)VBM>fy&OU1%`zJZ$>qxNE7~b-{acA9662eHJz-FUZL$nIT|y=@9k!v zgeJP^+<|W|T&_A2g+N&q+u|mSP$Don*Rfm+AtE z>1$sivA>g;XhW1B4m@I1^T9@elo^1cNA(ho60 zRAIso{gbNsN0bf0LK;*(o`r0e;5#(_nt0aX zQX7wJ1ireBPyL37E%Gi9fTn+`(R?y_2-DFfrkTST)FTXIGP)4xcjbqx{l37(fG4ba z3(2g>%74h7la=8Eg+JaBBUdiEJz@DHNAh{P5HsjAJ*@qvxdT1Lx=KYsLwBU@z%|lL zGz!Hh64XIu-m4QgIGI_f)&ClZc$LRHaZdr~NguAqu6cknw_4D~)v=1ekH6O>gZ3J9 zl(VB%-x5yg@AN@frWA}Fj=d?D{ABd~p_T`GeRHc2%!tqn&v&YH^| zK9Dqdd~RKTUWzfAv>X`Eg2EhN`P5k!I5HLVix%%}lkycg8~Y_ujj1*s!Lyd7xIfp~ zxH%e0g}$Ca{MEAH(Ip8k)4Q-)aXn`AY8RnOcp z1O_V3J`OB5iLeDd>JM`jF5xlQhcBB%SUdUs_VeG$}JrP^Q$#7qK4##Or`yZ)TMKPVz5j|HZMPBIW!=jgmYk>!iIUQpHTbd}rO4_WJjskhH6)NwxKLOwr4CQb^ zgt6cWh&SG#Yzsw{>XX`&=ezJJ`3Cfe5kfuy;H!py*QAUkW?cX9X=oyvoz3^-&}Hx`JP%OS(aBB9Wwhp@7| z$c>hr+k8sjIN=R>C%XkhQoe)X@_ZboV2H-v*Ge}&DokcMwpP&B8U*qRF1X_t4(-)Ztfc@k=xDyWKhrcRfj40tNSm2wo5!=f4Plq3iOFZzAo zx69{<=4Uu?P84t*MlzX3X${4Ju`%vX1i%C<7p5|{c!jbF%!Ee}*w!GY+41Jk(b6D+1MtUpLzfn^47VV~ykCM~pwt+#U2KJ?C*+t1!Z9w1V$PLJs- zohTks4Ck$6z4EJL{n=ijEuf~;6I?3Qe8%^4GN~sA+|EHL4tyHH9gF$Nbre=+#mCv~AB;YQ z+g*;x%#86AQ#M$kHwO7R<&PQECq_rH!}a;Jh2MecFBMWs&2*ohF0@qeO>0#nKQdqX&Me8Sd zLtY5t)~tFutede*xS=8KEC*%nmyB%bEmt zsmN&1Ht%BsghjlDK)#yDPKOhO4eBivF3Kn25H9e6+w5ZYCyhXjmA^zc97Czc_9Lt$ zIe(NA)qhmNezWnbQX~se2DKocS9h%5$iBJms><&OITCTk77|J(h&bB42s|6 zv%R->A~{zO;jWqcxSpo`TZOe`p&)(#fb;x_`i{N2_oG8wM_duqn` zkeZskJ#l=f=tB=#+7YjVifEzQT*4mIx)uOCq)DKQlM<1#ACkgacnK()1pbO@-EpAr z%{AahYWYc<9@hXS{M(^BuP@kGeHDFQR2e~cma8WTl6{F(FY#$_NqSJ9jg|QOmBDUW z^5m%zt0^Nk;B6^l^}X-*zVU+P5y<+2>Xc5UD}bnkx1Xt<+0gSdFWtB`XD1 zgeq2;KB)S{eh|U50G?9DIRwcF&4L^NdFT7;&V(>dAS#GcloVmQK?ohEV4DbY#32*` z33}o|RQFI!rtGeuYqruXotVQl!zS?-u2#;WpMNjp zGw&!A1lbAB0g=^r4vN|8diVg-D3C7pYVW<+@LhZui$Vu{5ahRn;Ni^h4%x_(0p#`= zo7?w@y`i*naoJZpAEZ&eV#wcAk`OMALxP@cHoF5-)Xz|am2Bmg4=otO&Le`uef{Dm zjQl}lCUq)%Kwli$af?)HV~48Y!-h;rI&8VcisR$B0vX~vZHx`4;7C!Gv86?KK)^s zzFe{$bc}$rSO>P>{m$Nb|fQv6OX6+5j4E|@%bZ;NI0LO9Ak3aonT7& zy!qs%ui>26@~VZ1^~CR!=6=?#jDLN=e?TpV76TQj$eJk@k)ji_Efcfl1;JAFZ-+Q; zma(6XeJdXTY_i{2>a#fSSEOcfQrH~|UobbOGeBP77o;x1_R%3Rp&A6SeoK(rMq$U+ z?p0E8Z!CY%Xtk|qcaUXOQgBoLWSFSU~Ny`s8zg%R+iL z@-8xSZy?OYOu#jSwXRlYiG^BCOE1Ys8W=JDsPl0whq`&18T%OvY%6*f75WbOhi#}0 z!;^Z;+1h$(kuJRaeO+W@r_{E(_uCqeA=UdjT*QT9_iRtUtIK z*;hHiKV9vYhYPHIFbsY)ui1k=-bE`Z;oo>|3?em3)k||=7-L~AI(WccR7B6O{f z!nYGk*>CCv>NUvrzYg)D?&uX@))D0fjbQ8<@z6WwcU(z(ZR!FYe+I3yUT{_O`XF?n zmp$MMi37M%Ovu(bg+Qqs4(r2V7%BILe6ew$SM658;* zJZ{7LPFBEua7r4ykvD(Ki)OwL_YTp@3P3iTbSnv=M;uF7k;i*yFM2KeTI(WB#sY)h zVs^B{`SOUawEDfZNRlh-p*a8(ONCkS*@rkeu&3*iCf097l*~(GmJwO!c0NN^ky)z1 zsJLiwm<%_gzrkePjTNZeRph_R4|#kR1#d>tmB&Y259Rl!!%FuUL|F%`FnSO1YoCTk z>gb7hxawh4F?z=AXq?+yz}IX6~2-$a6OW3q;afXe=*J< zbL<%9r!%M+$ONcWJtp4e&Gajy|DCzxwJ0hN zTd}2orf!rIk^gp3EutO~XhnU&EsskG&rwOz$M3u=3sUK;UyFyZ^ra)9RC7kG)ycNn z^W#Fcuoz58cLOgA3MC@yfk%zUEQYBUx}Bh;4~fbTtgb#b^hxk!6M(KVVyJXvGQOzn z3mA_7O@Vtlj<;MG7;o+~9KK<@wY+A2?-9rk$b$gn3Q#!!3$Dnf@=p8YN0U-c1IjW= z`&y&>!wMN$@iMxrSzq?y6OO9$^8^Rw9Ag+p;siu%DlSqy9L8W}J62qVy!7dNAa9|CS>Os&mQbS_tV;+A6`+GU zBpGJJA9e;RJ2*}7J3Wtv4MsSLM5BEwftvWyPA;&2GV^QS$#j>_LtUOiX29{4?mrMF zd;_w;F@urfOd|E&!lGs0H{i+@YXs_{>zkChS~UnDW6b;EVBvXDxr9$OH}454w8w97 zkiV~bJlqPzJ`5yT=CZZybK?0jU7xmWouU4`|Cr@i2Q&za4WTF#-@{WHDL#B>y;(jw zOHq4Vm+Y-TMk1W7YJ%nY5pQKbL_-;Qkj^;v2T+e#&;<4k-9CjIm?T?QxRl^`9)}M1 zDSN(K${VK(2Uz^DmW*W5_30Q|1N?9mx8z{GAN zENeuo=Y2U7B?t3Hz-+HOPN78JwexHvjDC^?wX39t9j0BZ1#3`$VIU4-DWkx+Ob z$Z_S7Aqrrz@M->9;_(4Whd6;FuXQi^(%pQhGSIzjrD(;X7#D-+v%k1T5B%mu9=Ak5 z^UL<>T{AR~C3M@TN_O-{2{)){NE^~F9&H}O+&~JzE+D(CM=y3S0^ssF5_GyR8x_jU zrIXbA<=1(_Z`~ll#p;aq_X`c=-S1CDF5taq3wlK(<3!eNIT8+eSRt5x43SclR;rh} z=G@QHt2(0jBm}UX-%_mISb769n*^)=KP&NHBh9y<36iO`2>BwpAPSjVp)10h3x z9;W98s-+)%=a#BW=`4trm36^5#V3+j6&N4j44Sv>rH|Ls^&8SgK}YDJKMz!`UIJ-5 z&EGh}D{C_PYY@kMWaowUBuL46d_^#jNKIMh!ypkKRr-?+qmGI}gMc6}wJ?zM*{?qN zy>vd%pJXi`YL-0GJu0D}eYkFMJUjLN7!qQV9?K$-@AaV@u-3X_UGJUJUq1Ijn{_M?zKikInAI^;cdlU}pZn3(+)}Ez!g^esu?5<}YHOAhh!c{9;HJvZ(ej>uYij_8OMff%rPb zO4eScR8z2Tw~|_~Qd@fir0IF1GEME$Lw;rl-&5tgV;Q2Tre{yY@$4%EPxvq!D~kz& zfCVKyZKonzME-;o1q}CRp$1MKVH-6akiSwmz@H{Mm5ZX!ZytX7#fq|Q;$2bZgQX8F znOD_cvyTikGc(8j9uka_k%z3=w<}xGSBRn`9C`)(?R`|z_L;wy!HH0n+4gDF)vJ@K zHAGlt0SK!>d99Tek(on;WpeyM^PW9SEmFxKgz!8E8XxvZh&g}#@kPy&-0h)@HP}i= z!w`E?$UWGM9CcExy#2QY?X$t7-)eHEiV zyIaaBg`>9G^h+l#yFT6rLR!Gqs+@!iRRI6(E+U-tn{Efw1TIkWkevQHqSRY9meHG# z4;`+TNGuaqr}{2ivyB0$z9VSs0j9@GJ{$!y81E*l0b1T$r5X%Y@!=8DI#i@9mJ0Xz z1&s+!9V4T$n`g|nN)VO*1pSA!C~z5Y*peeqe(T3GPKIB>FpQumM$NiEg4%Om>lp8K zaPLE>tMDSAdlAY347JU~P^e&^(36!gbzr+yDZesY_Px!~{w6izpRoVL&>ipJ+9TsU z6_oaUPjTTs--$mgt)ID1aX@*Xk_*8mThbdN1m$oA=k8kxM8x4&#~gq@tDT-!yo$8# zf1wHX0D^>mEQ3VGK%ftw$!O8xEhCvjf@FC2wsh)EM!e{PVW0n`XvNQj#Y$F5Y zL0(P39$(A0@>kPx4z2dOh>6ZR@}7Io&K2g{k>vMia)gt>GxNzF9-LM$_Sj`${l|W2 zB(Irca1p8IC6-=jnwo*aD}6qIAGaZ3@ghmC_^v*J*nnJwHka+Mdz(U6O@ujLP+5e( zmNdAR{`K{P^DDSUG6nY0-!`)SLK%4YvZ#4L@>snUzL9>8hC?gz=X*Qx+Nd+3dRT&L z-{3F&yT@u{bNebg|8(kJJYKs0LJak71vKP3b=3wd1cn6UZ%}^21o73eK!uQo8c-h) z6*Az6#$9yH0?{$v4*^QL07NfvOFeSn(w6FR*!7cNtw&u1PFhr4*bJ>pQXj-BsoK;b z^AfrlGkwO}biyJ@Ur?5BP+4p9pPhrn#oUF;rblBW$MB5^b_(FK$3DlYtcGVcIXhRG z+>Ffcvp%f(VL!)hFyp$$?4czU=Rc8rW9Ls?usvMm9MoiCkv`U#cvn{}V~?B5I@7&9 zA64^lx35|%piD_gwXYpCdom>PrH!9TQF`?@iN^~sun889F44>qkI`-WnbOZ+(~yIMAQP| zTe@2Vb@~}uQr2?jnE|^U?m@3{j`t2Skeu74N0hj_`f%3+j`N%Iql6d3xPn zvpgIj;&|llz8dIREZj$~%*)<1rJ(>|es2=t)*ukry|K1^!ongm9;)%q8qvY8Bti0p zb?4+uT^mx1^1AQM!zBz}#U{V_>Z@0jhv}-%1V7&~&g{!QFFd+FUj^lO zN7aP~;`u&LSNisg6>oBI8T`IYa2Fo3Hrw2h&x(0L%JbcDPciXvua7x3pByzFjPZ_EHP_+@gO+ z*<0>x;U;DP$9PqNVg(3DCfxDoMqG&_JP%Uq%u>dGorMye$R_!njq&G4Y~*tK@}ZU` zjQ^KC+BD1!G|k~`nePk<^Zc8lSK*L?Wei>c zPfF9=>OfAf7YRpbpal2&fyFQwrE`zi+~6-m;BVAOyY>oh#@Z?^DQMDy8n9{Y zdMiiReNwY}^JRAnj;`f?yKE*O7jrJBT)FXA>nD0jg7nZZH6TgF-C>Q&y&6-n+#+b> zwhI+lYS2ML(jaHnhNZi?qTFiu3<4e%EJ_An<)vXj=*_!$C;1sNYfcCtwfI5-WS$|L z=Xo9>3fH|a4fz@&tJfWzCz7}ySmk3NiS)UoAjSok1_~~$)Mo@!oeqmzV;3vnk?WWD z&Ui3`JU=Xyr0-dS>GZRjSrfzI?lqDG+c;6Y9>-IWUvg-M1Q#->lwkU&f5Im)_fdM` zft$ic;)?9M^-v&_7y6nI9P{)}GBlP`$dRyU1O-1x-~5q#fi6`!J@-zpul{4 zCVv|H(U1|w_&rH_iIRPG4%zMkeh3n@@axKNdwposR#px2lwcxWnTgq+FpP*PZxEb) z6r(yFOpAI4#J+7s?NLz}Y+Ph^ETmSpEZ$6JRGS3e$_l(1d&9H<=UvVH!F2GAKeM&@ zF3x8CDH)dkdtK)eEl3ne$mSNY*;or-dm7!$#YJE3wCgj6y_K-L$Xn-1SPQQt%o&1Y3I}b(8bJ)bRmC zSp7k(jVaM>s9d@u`aR8E8toZ1fhAO((cS2p8(mY( ziR(l}YVB2@{E0dfZ>(~HT)!F5+Pz;Sz{Fp6mBt3dawxB-j<{bOe-88U6JnW367gzc z_+Uj)M_Si7SPEbjy>!d<`|EIj)f)#&I(9KP0Ab`5fiZn1?s1poJ89BlT1Dxy>VJ2; zvDw~U#zXlcXbf0G8>X4ogMJ4oe-a1?AM{JJz4uKv_uJaKZ4a<-&skv$93BI+rbf02$OuT$S!&7~bZnAd z2NRSF^r&qh6on4S?fD={g?!34FHzFGKJ$@4NjVf}H+bHNKIz)2D3(*j8Ohr|H6jF{ z5tZ(%^(-&WZvtUgUWt{q!#O~*?B*f%z{B2{5Q}yLnPry|&vSpj zdEJS{JcL}L!OKRG_ze)U#+-ZLaq88_)FPQhKkPhd^H!;TQd7n2bIh8OqAiF<% zdCOqztO`0?vA|eKeqiGKjBU?Jzg63rB)3)vToCt{Sp2A`T4JXp250sO>4`5~&FiOx zfZ1iY)};U8JyB$?0JfaNNl!FQ^6{vD+_Xch!6Gul#a~&8xW}39k>< z_@f6yIKIyabF_>R8$#9jcM;o92Z}&CfAh5Et%DbB+qZcNv2`au%_DIw0sH$(HQecx zFpbMb^UFJ#Z?5(1lf2JsG;^FD2t_`4$58VkE|~oGc0yiUHoT+j?4Eh^muaRp7NQxZ zg@E#^Jq5pPUC@KbJ!*DgwB#6m_({j%>JfWIK&Id8^sO!*a@g@UEHy8WJ8zBM_Xv_D z=43Ejlyd;XJl9v}J7yI9n%xfG;l%uEw*!HE^JaOO5axvI1?Z(nh8Y}45#QJ*XvWC%|nNN8r7v75Q*6cpJFiI z0R$(JuU-PsiMk0cK9<#9dc z8ELKbbmPz?g657k3-Pk0 zNd7B(Pc+uVOJre^0ulEeQTCx`yT`!_sv7=Y?_+a(9?=3skMYBx;QB0pS4^R!0HkI}XC1VDu>e}w~)LZ;Uu z68vy|L0_uY8p5Ocz`M$LAr?-aVbqdo6N69-7R~MYKTmhfQ>x;;XChWCa1g`yv^`)B z_g}%UdeHlgi8luuZp6$#Oy%rCK~paXAtO=OX#Ww-7!~$|`f|{mmxWm+LYM<^Op(3m zNczP>;u{7h5B^ivgWdz-ib8FIhS3 zk0WK4&jFb*g*7LG6BD>PK04Qkcm^j>LLVKXYTK{-I=og`@N3~zay^XRrCy2PZ20?{ zYK1mZej(-cHzPHJDrv8bZ9tH-zFWf85=Z}{D&OAS6w43jYishQ#263EGJdX z!J3Ghu4&2>s*``{YE(NOrpfxpU$6@15lMz>_H{48@jEWa@35SrB%u}-@3qvT2Ir8o zBiWZCCC;H-Pns;6pxPfWo`T8hr#`3m@L*&J|DEs9lM<773*m9vaLUsg+e=WBO7Ts^s|fJj1mW$-Hw!^gwTe>x%Jlvl$b zsG#oHZ|LZ~FT=Wb^4DkE`UJf4Y>>33)&r!>P1?~~;68wtTRQR0CNED5$#m*(A9)GK z*{-LrN!-uPeNI~f@$%Wx3J_V9?epl9g-+RcFdu1FP-^u=lm+l>--3gzGnP^S>D)KO z?%sJ1_scxA6ngjy6%G2D`TH?Vn-{W&Fp3ZkgFo*8uDFrR0nGXtfx>@H@4 z4*jHkh~D90F;U(O-SM1b;8ugAE|dNG=zNpk?#HY2xx3cF&n{u<&{tA z1A|OIut4Mn6Ls9i(cO~Rj^z=|j`Ci-mMiD8WVisDRGm`h){c~T%U%&q?-E~z(f#qK z0pYpft$;G?hTA91%R#kP6049!&ewYC6^~79KO=9)$ppADH@jam{~KCaqSBHkg4P7nX$shyA1^c=`9O6tHqt&WaTvLEF9yz7@1ibQ%$J5Vc}mwf)Y zcOP>Kd7%}WBbR0b_MwY95)1AXL6R$H_2UkAN8si4@9eYb+&GZV$OpJoYG zVTb5p#)Ng;$!jD6uGexbr;v4&43qlaoBC4LSnNDev5bq841O=|G$@Tj^Aq(Nk~<2( z&!js2p2JuDyn4@bwnM&0KeMcx+`jc;x{fy(AJ5=O!(?AnMCv!Zoo(yJQ?;0MJ>u><9f0nbQvf)hWTCWP=MA4i!Mu0*vN z>7C1XpyazBbLwqh{M9a}U(3SJOjT7q6xeHx-KgNc>7a|^9TLboa00?ig<>CWTW-Jb zulJ;1&{##JydNMmb_n)^t)D+@mY}3IP6ew|ST19%=iLjtW(EX*s1wxW@q>FqEGBH= z21T$@tv@-^b3;~799G;_c&NT91kN?;-EBVahZV(n!LS%$gCod0SB-z)PG>kOg7#vV znqZv)M+F*BF8|tBG-DZbQywC3Nh*lSI#m%pZNDi;q{VY25@8`!hBBkIPV_zsz4rE( zAy^)=1>oG}5g9P2Gg!&MozTuNcyYmy9&TMuG>lD{NDB*BIQ75od&PE)!W)ddvR{6h za&wVxY4p!40j2-hLM;*qN)2dE>8YW6!(eAVG2R89E(E^~w^b&OWQGD)mwF*8_{AvZ zky#L)slA-rV_)r=0fnNk8r7CKhQAbkBJ<3528Qb{zq)w$Y%Y5>IHSqFmVse{M5big zL#FhaS}Ge00;^y#!HUYjRGQ9?t4xo?Q4nE`>_3QD!6PZea9kPc^-jK|1sf z?1Yz!^ET|IOAO?eeQDR2LDr5*1n9k%IQiSTV9sU6EXWVY?M`C?gSvvOzo&H!i6T~YB08#-OZ_pjYWU_BjnK76pTJ#>lBmeR ziDgSvD8v;oZX)4|BlDm(v12L7vbv6y1p(Z80iZUqD-kVm&)!Ul>K4$Ztqrn5YvpJeuJQAAIzj$B2IxR&%cKTe9uT-1$0E9axjFNF zw?z9}aykFl<*pK>0%$>Jc&o5|l$nv0fy=K(q+Ku@+;s?0r1Dcr<}Wak^@&>WRy0b- z8M%vD`->Xi4*-$$IBW@YIF#3ux1}ZxaY%j%tp||sl*^cnicIgOUXOQj# zI8?r)lQejp4Oy2^F_xUxoJKu(6cq2atP0=bK`k0a4Hyczc;ZV!oVKhU8rQ+-vkC+G z3bSfa68sJ*zZdoc-*OsgWGwU|z@YXRklMiNI*id~MJEJLfvtdo2qa-nt&`_;mfo(E z3N053R60GoFuq5~5Hl;fNi($oE8NR~`)Bla^2ntQtUVHKbp39I!cFgc`ufmpE6;RAjpkQap%B6^jo+#J!&fR0j!xNhN<6hMK&BR9!)w3%t19Eh zD(OrnK`j)u8Wr7DfH2^O)eQIH~b(kj0Q*?D#a%T%j+(zkjJm z`@bdwi&#oQgAS-zSw#D--*pOe*=?)~K81Ms!YG-+`udSLM{Ehrrl22Nm)AZRiq~?V zPw>ZfWvJKhN+;b4>UjUNDCp2J1PI4D9~#ib5S1|Ef!9n)*X7dw_5#NlFYpv5)0-^M zSjtrB5eg35>kN9+VWtRm%1q%P**8!}J%Osw*|)&(!S0JZhe$vSq2+`6mYsjkeGj^@ zbew?qxev@pwSuuTAZV@;Q8+u^kIInz+0|mCTr7zT8)1Y%fn#{8 zc19Nf-XrxLvk>i~TlE7+Zf9meikfYozSxn?Kz5)D^R4VTyes$-{xE-rEh9SEk&?Jj;ca$ zs@uq4uO|l#eY($m*gapKYJ<)H}VP-=vIgFEwSV?rSmOtRczM3lL5oK+- z!dFG*k`k0*d^5S52liT!z)c`}y${%s?lbPV!v9WALqn+Xs{l#{^j6}6(Ft1jmRP&r z(tO8y|D4uuGRW`wc72kjs`=0Jzyymc`HtZs&Ec9$Gzmw_Kb9gf9R z3E!D{8|?)V9_ydABYNX#Xkj$q`d{)h(lk(!ueRo==KMQ|lIG;4w!ggr?VE?V2t>r| zA@LiM4Ec7s&`0z&mp9U5?-_r$ZGJB+_OwU7Y4iG8k$%Xl#ZX_b8r_Nwal2qZf{=x$ z^(P(vVuA4J%Nm8+4QVsyM0FC>z*Bp&-`jpwgbJ44nUiFo=k}E(IfE7sfL79vEA4Ft z#h0oUT!}>fb;TK=;VWDJyhv5gsk>yIW&&5TC5w5mdHIkKRz3otH^ph&xI1Q?P8D^ z&Zr|B_aBzZOc0GpDvKoYy>;knc87tuAxZVvZjBi1+eTFEUJR2^^>qAx+db9!a^sL& z2%XDoe@7S8E#SF3jRr^VzbzCif`>eg5w7Id>)!MgD3ZE(gmkcx*)%;Ned*7V45=n; zH;90969jgQ-h6J)HTC8969gTWB@3!@3-2n+bC;v+`)6VWCqNRlJNEKyV3A-nBiO}v zj`;G;vz+^4I&NB(8TpcfgL5Kc^nyFAr9W2q$@lll@81=(R4Dh6ea~_B*uu^?}q0$r*4)s7c=bp#-e71xb|B7jlA#yE4k-zW(G96TLaq(6c4{S4d}{)@K)=P+&6p z;zM8s+cJ-~vJuWD^WP4vKJE{Yaa?WjbJ%{!Ma8p_aWHtpi_EG)F)7c85{$^>xIPo) zO^h;Kzdqvj;#E)dEXn}jzR$u(ny6&>4Y%K0qpbb9EWg_jmLg!H z3FH8tH?+57T=)i}1UvuKyfW$YCa=%G;AH)FTfF8IU+m}A26Gff1AVr>Gz6N22htNr zka(L<*>D`;1dVfUJf`O!J_~k|@DZt@E|%J-r#MJh@8C(yw%;+|tX$UfH(D+Nz`dw$4J$s)l$6oMMr=!UPDMQIs;QTp*Rzy z%HuLaYjJmJaFU${b(8Am`p)G-f*T{gAYwU6#T7()BwsH6MTZ5I{F1bf;P+%Ug936A zI3nM5`vfnBC|f$HEl$qn`*T~kgsq-Rl4C_ZgzJRdlk7RUr|a^utGX_9xyqJ(zPJ&^ z1;-U6tF4Au8CF4KAP014k8#V=-<;%qR_)GMeOL-GemTKSEDWt(RcJoQmNNSBV zGF&m^1irJ@@9*#iBdJpkVU((3Fje12;H@;um6is~GMdX^at^yl58boi>+W|G6nIDI z&7$W@XM2fG$n=+m-ht*fiQt)11b1tqV)y{N=>;9A!^L`x_?QTfwKH^Ws+vT%A1QnW zogylr+ESj!WRB06P~Zlvc6ZzZPn+AfA%~lM=GG;kM){=2OHC`wM?g3Db?;Qc``S{~ z+=6D(ZMTp{elLp|kIHA<uK%b2*B09P4t85~yuK)4?iYtawpKEj{_jcLi93Qh z_Pdj-UE$gUrIRpJwOt;G__|FGZvXJUt^Z!YHKqH)r@1n;?nBG+9UPxjkXmLbb>O|f z3{+qJe0aYHQV3Q6_Yr1=y#P%XbXIx5jvj?_h*J4;*8S+?vK9!@b0`lKtZ2|Xp`^vAU`ksst^1Z6dED0-7@q2=5ofKbMQBZfd8`u5! zO8}ejAhAM}`w)T6d9Of^g}m37aw=*AoUHY}PMb-7E zTeHt!s;HdCnH<35bJrvqd;Ax(Jiq z=&ixqsfOg-7UOVKM@-pxd4??w1wq{S@34sP=l2$}%km8B+k8rw5ucHE0~VFDHrJu5 ztt8WU2N&dChKT|BsoXvr74+=rbhl9^5P5QZh*6CNB{>Tf?1HFqx(rwG8M<%hkL4mB zVfDz>c)uQs>y@n6#yM6I2PkJ9q`8}C=nvpGYGp4Yxbk;Ke4l)*(QXz?=EqU_m03W_ zY)zJavoiL;S(hNOYY~JYUq=ANUtlY91v}Ca+7|5rwccR(>{AzzhN1CYE+uUVLgR)E z_^h{g2ci`f^5<^e!l~`R`La_H7TqzDX!bWeF)OpoJ~bN~IJ9kArEr3-jF>`|oZ}95 z9W(y+RUYczy1~(F(bJnno)|!<{3{~lOtW%QBG45Wl6ZkOuSJOXA>+`4suonm7$YzA zvw<0&9gh4laJtBQYrt*8Cn6=(`~$CLY(~BjtCrTZPg0vuQ}Ilja*d*Zn1vaS1VGSq zR%fWyY&k{P1PnzGBvPLVXkf6w*yybVw6rkon;!DaW%6~oGp;0TM|X1;`KI(tdt$XT z{!JXgIC91$kJ08YvC44pvb8#lH$TK|Jlo`qP071u{0#J5@?4;1W;DW(K{EqhJw&Fo z2Obl2dTgj^Wd~@~WBf@$C^YpY(m(F)F3y28T)R#W_Qk_e3g|EE(6u(o8#Wg7f75kP>PiVij%5VmUyx(cX+GcwqJzvbo!e{I7v@>mTF$wYqn2-9Q3%4S-? zuld`XoRCnD?51vvU7?{T*X!@<4VB*z8dt7X#r$mnD^-K*O&<3%o$TAudWs^DFAy^s z%n7T$W2(YoF$1a9rRvUnIXGp5O68rR-zAz2@Gml)3ATdbz1E7zj|X33KllxcdQJau zChr&!JroT;`UhWRz?2@rj5pnm%7k*g%LOT?S|>sNy?ePIzA)UOQQTG+JXP&;4GUD` zE)}t5)u%2n1mWIgbxRmXAVz~vp!Idsy)#B~;ADOvuWfS6F*8U0jzPbIo#M!t zklFYS^2%m$OcwSNIU{aZ9JX46ysRMZQU@L}aurPGaULJYYe*sRcRO&KvqnZps?D~; z?w;a`;q)F|JRR=rw>Igzp(V%Hx*^=G9rj(_jEd**F!~N$A3zka&zbqh@$W#iFnF=pn;f;_0n<S zb^0?yCj0dR3f!?ofhgc22LLICES2;DUDl7$8mUbUL^vTGF1%qE!mZ|-JT8xmo1pJ5 zizglY*e$tx9W*|bj_v)KKvu?az1>!|Z#rU1pT#FxP%lyg0qa&`*8O> z(FJ+gsz;3b&S}DNA%{02Fu@adrnft!N;5KwszO_p2T0+|>)?wZ2`6j9q(WFo*?qNQ z2o4A2Qfe>aa3jT(6X}q=o^CFz*2}+ zQ-@jOaN0F#?EUB$^34He9sz3H>~PqBH6gBrFGao=;m~z^d3<5QKs1Bek<^oTKdGkm zI;?vSSYJWI2}^X0w?sIHGjW~Rq7=K~IX_=g9JcW=Q_=ANrn-2mTucW^MW)}AY>@*p zf3@t(_lFEz(ANsVH%Q3U0KFM6jhGQNfGgf{hE2%c0*wKp>-PcAk z&2qJA`jMZUWf<-UGfRGnR`Ont6i|Z;4*a>`H~Lx`+6xdVAn(El#$i^4j?)S9RhPfo z8;mv(~J8XFAxn!a4iwUG>%UhKEc%$m01W8HC*!eMa7q+dZ;QJ+A{WvXCmHhQZMq z$(L!?bk7s;F5|-caXum~C8?a(#kYd4-#~2qe!us-^j=lVcw}gp81cQgU$Wda1_U9{ z!4l`b`u2(Z{?Nd16>x9AdnNZef(2j-8=;fN{mHy0L@MieH30_~)K;kGc?**56!!^n zz}kT~cAW@r+DNy%?=72|-G1PLF1jVLCje(V?bX$k5hF&er3QD_lj%Q05pPoNmIGtSw27J ztxlz%Y5PX}yM{&%936yXoHv2+Gdu0OVV+AE<;qHTRD(1)Ev+WF6>3K*$WBh%M^%v##;^DGd$%|oi7 zyYF@+(y4>5+_E87)P%$No2<&@T3{Q;|C)(7*zQ@R?5iB5m~F(> z^I_KJsCc&b+v=4KWUJg4QSd>xwot{9qI13C8xW_1SjaCHlY!krM9cqAX;_efw*P9p zcky1va2~yj(f2Pxq$)~TMf)-c#M>G2Gaj7zD{mTuz3j#Hg^@${R@_a)@ucA6-}bM# zd9&ogZ~Rq3w7&wIMx0iKTaJ)awS0)y+@JfR%n`I3xS~0k62z+hcbB(BI#foN4_!1M zlAdv+iFr$s$*l>gy+6-Q`+dj}v#_~X!0WoY3kJ}*iQmJfdH7x+^|>|b`5O#K0lGd> z6KeNyoWcgWK-dRctV|#>kR4i)54pT810HS(0g9e>Y*w)eBdyW?G2~E}ir!Iy5&F+k zc=O-5#sed$70DZ0z`o|*u=9=06{3+#*ayz z|Hgaya@y@~1X7FosHp9@mkn9f)d&4+_$LC6JW2+!ru#-=VyqHw4`)A|-P2a_2J1acw|@y2|LmMtiaZpgv8R8rosS-ItpnhM zvi@1NGoN6MMc_%ScriL72>&W*puTub(Of%JIQvY#T6O50g7+dBTY2AEp}{ulv*+mV zwuSon2%LCS*%rtZg8Ob9zdl(r9_vKJkeW|tsrelR+DL)+s!t|5g22)?Y6DpE1vw)1 zmF@P!4E``bAZ5w>sVy+@2?BeS7{tooQwU^WZ2C}dF@@{_ec5SSCs*{CQ39UuNgPLr zxB*}P098_GXiozzrFPFPY!3PNo`T}KUj|XT;E6dmQ3r`SJPX1h>TOJbH{rGtEY$7% z_}T%F>jb+qgfZOlcM+U$$TYvXKo{=Ck3zWy;9yIqepX)S=W;Cb7t$LMW`pmS=whc4eVZ%RRl@aXO3O|EG?s_ z%iWVQS|E5&+3!Fy@{Z8e$1VUEd(2*j=;`|poCLle&&49c9H*8>qjWY%YW67|$X+k) za|TlkwTHbKD!SOdbTvzW0W5WBE<2c4`1KfS<9j*1}Z+8Vp(}?wu9^z?U#zrRsG@!fZ6+m zDU0H1z5#X!%8orsB@TIEbi;NF3FVP8W9j}*)wPHSL@_W~zlo8(D0)_GEI1tiou57h z!UZ%0Y~FY`f+C5mgGiq0+PV*;t06mJmQWaxcVhK$Tz6m-de$mkVWLysv55;Mm(=@P z8JbEtnjSj>Pp%FTiK_RHYhb5KU%0f`X1u#&1fqCawYMG9u+&FTI^8Fftx>Tzr5 z@7>>*aPPLn$`&5ucX4My5kO$fiZmWa;j{4!>l8$a`>vq9YcFph3*s%27P?3n@LQ@L z)y}FCRRzTvelA0~$Lp6ZF6b%5;rjwSd~pT`6>Xv@2cvR7EZ9Pp@z=eYf{iI&>h}bS z{r)cJWAv&3-Kx|O*nt*x##t$|jf^zXV;rq4s3e2L7eXuAO|)<=qEm$J( zn!cxFC(Cc-RRSX1O`JD(6oVdUZOqIA2$wsM!A;G;b;qR@$ks@DAxIY5a>cao=Uosc z>ToQCxTEn=n1c4Z@pZra7UP4I_`TA^b|iKcVvU4+x_%TT`Ay|zL@-a9)eW}Z;{@;X z^#i7vog%dHx3|6UXHsTYDf|(+^XErIG$tsj`0!<#JOMDDdp{o@N9xbgeuUkTd;sH( zF0QgMUZmofRsLxP18HC!Q`^jXJrdvUYIxk|B%D>!;!apnSlEx{hrrrvVYr3yW1+x2-4*jg zFa&M%I}=BXckoprC61(G`9@t6Fzn?t+GGTy7OG77Sc_J#wq(>xvZXY&U zt6X~m;nRv6sE4NY#QJW1S0f&gVZC#XzOBTAc!~Lnh8X!8IT<(Ve51!Cy;A{Ttz+Mb zdr9&6482~Ta4n>X#WU61#o5fXm|o%@oau{BjwD1qt}x`+g9KIt$AQ?iKGq*G7v6dA ze`B9yHVVZTE5y$PabFSW%R>K9zTfX4bEIrbfL z2cdvW2@jAaaGC%Cl)FiOz`s!;yLQ1L@c*G(OR}l&-F}^r$>5~2%+(Oj5lJ;>Bn3*NSWa= z1R-=^MK)gM5ROG!yS5k3q}Z0rq`5+KBqUyUx30*e&D&=o8cA?P)f^g5M$yWx4RzUs z^%ZGyD?D$ldWP->fb++fx)IZsMIXD4?26^UNUK-#;a{55VHZ>B6y#rZdv@Fra=rBD z^ykCP{3510b2Ze9?U(oN$F-Z?=dC+kUG{_dVBw$(iokYw;kArn-q3!^&E|syR?ZdR z9}Z`On72dOy-DY1JcqsJM>3z9ws{vJ;a<>?xxzkwVv%r$;sf2co&t44Ru*hhz4jX- zS&|h{_|3_>zis{l#jTI3LsALA0c2ZG3f2Ex_4WST1xD%(BdsA9C=xz zd>~u^A^Tjz_AE8$5sdT;1J7wPs+uCZEmIwsw(c1c8^Q4yAs4y%>FibBrxZmoIWpV8 z$_;IVVAZZKWniH_=-1^ipIcqJ9%LbG(D-!iuBb&Y3Kv=LzVFGu5wn&Os#~NBo>Agb z2ZXhalw^;`7we}597mL)$^7tPd%~{9M+Kqo>SsL z_AMXGpde*5x;wVS_&r=`OF83LhgC7-O$Zy_Q}ogYedOlOMRp^vWSY_<&rvAO!t}6Q zojV{P(5`Awb8Cle3PC6ow5J6KYcNT{-2{gX$vxt7H-eGk>&t$LkEP54H<&!k|jU-R4R4@|Ffg7Cyf7is&9om*_610}UWtLcO`%Zc{T( zvrtF3b$Hhx1n&?zu&i?>WOLff$aoFt15KDr0&?;CLNs#CJ)ZCT&<^`Mt1R@ps}G@g zA6ZwZs!+D}50}RW?=SqJ4PiS$nEj(mi68lEueRW_rW@Tcm;bzwg6{LK);(UeV}lDy z3kbWgl@UDm{mwKm1@Y6TP4B5RTim{vlJid`=4T&Z6GFR^N zcNc*ZfsZhpc;RN0%`7Wnri18%ySq1Uj9KQel5Nc6OkP9jCXr8~f~05te#=vzZ)(Fd zm;`ThaPE%GWv^V_>q6my#PxoMfhm&Z}bI;$KL2#`j24 z3Hxf35IVnFVkH18j87l@@(mU#VAf$(d;{s1$pu&_iokQkq(1q!ms}}S&W_Fjg@&**{FgvjW7c}beLG%uETZEXBhqK6Q zh)l50xKnEuKZygYo!DxO}<*gmw>AB{!}7bL;ZP;|a!Ztw=yGgTLdL2Q?92XlTG zS#ywF#hOIU#vA$I6O{t0wus|4JS!aQH)61zFcQA zqNd+MVPfV~+hfJFTbu-c&Q8h=A}8 zoe1@Q%luUxo!iAuZZ0GD7E!~t1S?fOzR}I{Lwyme|-Gupd%=~*TsPFQ$w{E3pSLmoNQ_W$0dr5%eJ_+weFnf7^anc3+8%%OnehkYM3#o?XhP57A)&l+6$(eRSBQehJII}F!vBuzvsuyFqVTXMX33}Q5Sk*~Q z0|WRXHJtHi6c+$4fkXkzw1HpbrdJ5)K4eXFs_a^`Ls&mV_EaO%2G2w*Kl3UO(+(y@ zD&f6up`DqYq#$k+++Z9c93JyA>!28h?;^d~z*%py&G=FdyU9sOVV%|lIvbk(j%_8} z*?@%pK2PoWxnD%L{0?`P*7*s_Mb8>km=I45$vht+(o8M2vR`EWv0R3&ysw7zs5KZN z{5g*I$Xe84a3JY1PK?ZV9zvd1#_M~;E5AB|zh!V>Q zQ2cRlU^Yl}gi4Dk@B|>|(glPa^hf%72StM;G!<>NIh*t7P}OZLD&`7<9G@93r89Ky zBk_SSc@qevC(0T*qH%0I>J6IP?1LijI;4V7(jJ))gDps>eG&AK zC>b#YE^To?ZXNXZ-f1{U(F(PA-=NA0KvOnlo`Bk>UtEOR9q-L58LceM>NMZ3jFcq=)Fx|McCb*5NvXIjtCs5U@z)@<2{t5r#c^8mc%w!t)(Z) z;dcT2w+XrkIDcR&2sI)eZ&^dEikXNFoXzTaD-aZ{A7{s^9O}$f2tW9)?m>&7vu`l% zioBc0+^C42IrOY{y9$fQSjO(Z9EMHloB2S75tSH?tQC@+fMG*_#fByHa3xc9M&ICr zcyl8+#t4VRnCg@?fyeFlEX(69lV`jDN-EWew!_~F!HiL4CowG zj|SW6ecG3*%ul}%szbu3;v{Ulr(`7Ux4g=uSym_|*P{E2k(t@Se60II1UkCk!54}Y zboY5Uo^HT|=4S67T&e-v3mPhfL84rbfh}1{mxPC@{345MbVB%2|wK?25p)koe)j=aV@Y5KYvF4}5z&dPX&% z>SYP}saQ44;lTH$Kvhx%iK-1$jJLfYECDagXO&d8z;MQ%;SQcM(E*|Cb%+emG2zn` zQ%7>Lc7%pgCy?^14rp7vsjqSGNh#urDlHd{SlOEo-_MCfJK)k);@NEh$qH}(P zDB?x2ry3Lq;_IH5;=3|Zy{=JjxJBo^f7S3n`Y=gc?^9sf(B?xRCD?A)ci^KT+NXd0 zZldcWOIuJ;Mr1(50G0d*+sFRU#(6&GY4G9-p;DmZ!7r6x(B8AW54e1+`p|myB^&OF zTL_cJ56iUe$CHP6KH8Kk&#v*+2pw7-p74^%Wa+;Xm3|AQ9&k80ubK=vO~@|>)Zw3Z zeS-JZa8VG}i0ZCRWRek&OYoNN1;x_=PC8{FlBf>(5{lSg>95~E4mzdl}6<dt#*g6^9}&ytymD*9>4N=E zNTS7ASL1ENjE8lbY+(LIw^Q06s3lIL4(=nW{`6~X za%JKvA=QYQwvl5R?+*G5{9hf!eEfGmZnua`##hWZ`~R}j_2bNML5*7E?|$}Ol}B|q zR<2SfU4;9`WJAAS$xVfl~oIrp-QkZR>kq} zV{BxH4%}}D(}=x1v=@5pN-6|8Gh~%zAp=aANaui zhTI-j^Y1{(eR;&&_j1|XjZUk7$WU+Hn1nUX?JW{9}6WtX*pB(T;Q z3Ot^68Jp6DxA(FrRGvzM`R4#VD&lCEQRW3fq6TnC@tP~CTvF|=8HlZOmdnShA{l|r zva$e#F9#jtC&45>MAo`mQyzcTjEeQSz{Hs;T7?ifsOgb1ddlJ)M4N(5W^AjqqeiD` zHXjT~O#a)Tx4rphTx5Sih|I?OeC}X(d*P@s-TmHp9L!|$U)|+KfXgNBN$=rXB4q??$O$E%hdsd0#Utzm#!y^cFnsqo`Gf zRuAc3;mfoj(3TDcf?0Pil7G*?XFeJ$$z^kx0Xm4^4%xGmwx4s@i5wFAPmBRMmbo14 zKwG$;jpyk#oB+lNT#p{-Y-(yw1M@dHg*|8L4tcIcIOyZdr&r|o;(@6!BJ-ImLmTyw>#27cs5B&Xsy zzkZ)5YZDJ9?~I5Qq^pq&blR>HZ#p8|8{c_BqL^*I>#|fJfoT&mZXeM1^fppzKa$63dWq6G~B79{^H;1>5CXW%o=W+wjJ@xzvz8A`Q&dQm*5lK`tHlS-2w52 z0ohYpa7#wcdLaA|H_kyjq(XjW%p*u=&fTd6H^ZL9L9JXx-75#y@#-G$cbkd2`2x52 z9`oHb35~3d<15pPg^5PTP*8ovD~R#{cYTNA2Y!e8;il5(^>32TnY6{03Rfxpju{>^EvsjhOg zZ;L7p|4sabHUk+A#u0FVGb0DIG++g_1MWH9<@2IK$i+pSpN0&<@&p2|67tnykd!cW zgky52GQ@9tG%^7ymgc}^;3<;?LR^5~MWjmK9WC6}#UXFG z@Ohk;ykus!e1)mXI!s$M?;F=kQ*(IC8c( zpl>E5ZNN$T8;V9^fH=ya!+jB+{X^D}fMg>ek525V5+H{q0-;fA7jYF?m2>r5;~-|E zc~6vlyPh}h$ivgA`HM$;h78touIhV=O|gri87x(Neu?Sj2_Fw{p-W z6kgn^XneUR6uv$ zX%*?p`j#df%w!ze%qZokuu9wg#4e(%L6Ou6OF^~Q;{!gVX;;WI#<{3=pluHo3$u40 z$a7Nd^}3eNHSbOuBYKmWL?VFucAwFw7s3k|6MO|JF9 z;d&-06816+rXg06{CXr_I*0J~S&FIr(Skm?AB|XH6IvRFo^Cnb*75Tq3#)rDarfY6 zha5F~4{T8G>O78WQCI4UP&h=)w{N60SN1_E9Mns0TTKZv!B2lBK}E^-WY{Y*{=oBw zSow822H4C`$A~J!o7QmXj62+*y65pJqaU8&$X{V>d+=mCgAkVz< zTkMt*ylO5q`g*_i#0BD!SNyr!?$>!D^g-KiI|nO@C_5L1m%MN7+$!`7?#8S6*u?(n zNhf}$$h8SU@-$RyC!_=9Xa-9c|4>{_#YN+w(2&UNVvOrj^6v?8(yfG7I)&FbL2K7@ z_Y;)imDdFj82z}hCwK@TB5~}_L8==N*0t|4iz0{h=~n`A7FDlz;T)}(){zM)7N%}? zK}MjI)17ucmKl9Jvm6h9o$rE=!6~4*5%JVrPbZlzTq^jf2QEq$!2?38yT0B5J#7|s zM~BD3n!DJiCc1T~$$*@$4!J~74~FBI8pb6C2t`ujc@if2dr_QQ`j8~maIHf2muuo* z26T_y7+=gf9p;wnu{(X9>q31P#Hh-;%D<7P!XS?j^(^}#nC^a*Dha(fw#oDATR+W| zL|E*;Q8K?_FAVXvKPdsv-o*)1x&{{H9+$n5pt|4{?0qEc-zY41{36Q_xV6s;kx&=+ zQTv-}3$>g|o<%bkd2olMyo)|pFUm=Y1Khq}G!RdZTLD)(R1X+_v$bK@cYwzl%logp zS|#}n%pJqf<)`_3e$K`Y1>=nxx5LV?)_zvuPmQT_H2Bd zK3XN52>T_60LOGf-8I`B239Ih9xt+KER3b?#~3)RKG;Io<)E{R|7%H)6f?caYUKDl zoz5kP6Rv3XCJ@x;fyHz7hRuc$tf;=9q^!Y}q=fsts{DOPd^llk+|EAO_i>^5FyDw^ zRmH?!)R|GZJPKvge&;KlL^G!NVyET8MsmA;A80K3$g5f^R|VW2DgbC5G_IMoVh4jj%X7 z#Tl}~lcl$C1sZSO#iiZNeA@XHZ!L)c%vKSMObz)ajPr{cvLbI}EH6!>stz8j=K zq#UN8!rB21iR3j?+_k)zkiDz>Jgp95avAVD=&Uahg$i6p$ctuQEsvrZ?7MTMvqrk= zwbrO&j%SWccPK@pm1GvKj{379U$p9Jy>-qdeDBBj@bS7JrZ^^}pZ}Uo0wyt8uPtk! zNcrjXh@Qwr7Z22!AnJxGfc>%`9n*mVBH<261L1dr%4cMg%2;z`=_truU(FNsU0P0E zRK@icw9{;l5ya~PF`c@6x5+>&8&M#(=PaZ zjRNTk5F}Ns@4_Mi3DMPVf6wL9bK1D61%~@Mi>lvyfONQcxP%c8VV&ei(9(-*=;h(H zEc2s0gx-qJ;b(+55e2HYBCyu7x^tM{An4Gvy}2sTxJ1xa_`WMsiRoN2HbH>pno@vPI=E`XUOcW4-+(^f5AZ&GNwGzGfk( z#21>J<$HduNA??BYcc^!PSj+>0Dtb0LE?aiPRQouwJ6fs9^8fg44@Ej3>8Yt*nXmA zgWr9kOwv3@2OOLGanqg$xSSoRK`=CW5zsJe zm@+!Gc+gX<0Aim|Kr&;O;lzq-ps_`$MfORN?Mi}^C)6mV5k-J6y!XPw?Jpusg#8_( zHRwP8{?#XOqge3T?(LYr`(LiL8st9wvw2~eW{o$$Pte)FiQj$SIc9ct7zO(yMw4h=JxIBHAzv!ptnhCY*-SFvsUO?VXTPp`GF1Q=obxId56=u?;77!w z-oaqYa|h+2nD}gOYvNJ1Zx0wYqcjwd`rLb&fwXZ9J;{@BSrM@*ec<^va84xgRL4B_ z1r;%?-R1#30)&{%zUhK8S%|&(b?Ecp4cWRiR3gBGZ+GgX8Cq1x?p#DyLy=ETWPaJ6 z;V7$_ewQH%R5pWk8`y7$?qMS?=(id-&Ih0e8#M-tx3`?@V>ZGPg-}&Ig{T=`pjTmJ z5+NE6izL2me59HD=E=O@l{n(@5dxfcG7DteChZHu-EonB`>T>scNV$2x6i@BfJ#Nn zghJ*I(k?$z-iG^ItSFVesM+BCtoI)BrGM$8bOopqoPXOOo&NsVkt)!Ji%jjykNpbR zpXf1+IW5?#I0m370Hm(TD>yWy47kNxb9-$T;2)Lv$Srq*W?nq-0iynv4yQI;P(~j zer^s>B=lF>xPJT(dsO@e{d}H!z#T)$T@7&uT@S0Zq@naZ%iVQ!ui0pm>e?eoh1GpM z!TEJBqu4tG;pde@tL64P3zs0ZqDn`>PfRjMOWUamj{u;adIC>{I&VgJAih__C+D?> z-zWe>>F1r!&&OWfNu7smD?YUAuVK`p-}-sloa1LcU^ORPA2__*J^Ae&`bd&F#RH0@ z=Ok7~D%qj^QK2+PKy<`9!@Jnf;VGTT+pCq-cSj)};B16TS)-aXDPl5&66>3Dj2``j zY3}XGcyv1HutUsvCbC>;PZ!mH4_pul@mB~`Ps@@#Bb{65MbG!a8D2XxMf5w5J|pC_ z)4j)DWQM58J~2VXzX=o86m0}9V5FpNAHw-Bhm8FCX;x{$>EW9pDK*pu!54Vfm52@N zyPw#Sp4H`vczs`u%-#~gK8kNh%lHa+_5KqdFwEcw#WIpmp^83WD*k;QzvMtk2VzyO zURC!J6hZjXd;szu@vR-d%c|Qt%CxdUQhEf&aKC@PD`H^(du6~_3*%;M`fd~6Z*%0R zRE{tn+2erFjd_O>6>?M{`8|^FlI&7 zARc+fhQe+s3Vrd@g_sn9Ju3>`T2TCcAwoH+r%QhK19Tb?gMo-B2C)QqSh3DN!uRfA zI_!1HL&N<7Ht^F7yZEPl%ieV-)oyQRaR2J5Di>FMY1C`1y)%8g}2n`^0;^rL> z%hTsLr$O+z6jhV7XU4Ix932TA3E+isWDqbZCeiz;1sp@YKkp+&(Fe^x^J}NOI1X#V zN{1lab8sgWlLA7i8r@zraT%z{cQ5K4go|lTD?_;LzwgW+u6lY{Y$JA$zF@`vWt}cZ z8yMz?09BJe$A?@#dC7RoHQRVZ}Pp<8kJ84^yEs zsp0p!aD8uY|InUs+RqK&UDz$II?2b%!-lWJ74~C3<-iS1M4&wM<70gU^>qemqPV|X z(s3jBpAguHiWpYpUDj1o>XsWxmGcs%wA6N;*af1 z!9(GMe{t)@E)8aN9FW{^AYE(qXzs<+S0!w$p~!8FHCg{d+_EzZcHOqNu?Jwh@x9Y< z4>O2Pf92on$kAnSdFhgF69L8&p-upse{ERe!5nS|wl-ElLmy`bfuG1jLS}mlPKkxj zzGW<@1uH5$ull?u<$|<5jT-Mps3AeDynEFrpe^(l2ZOoz#ag|vIL}4SBK2bbjRbtF z>?QB~BPd4>M`-XZvC-(ritl2Jo2{9>RGZpJS*)?ZO)2CLM8zf2 z=Sj9NFh}n-@}ld*#q?}VGm4LjytweheO|pppqdZ(h!$>tB&5D1jII~oDUJ?eKhZF1 z#iSKzpk3dAUzhy=d*7KH+aiE{n>G(=jz@Rg2jlx+}hVyyI!IKlAI0+fmK{ufWqbDAa63NtVc1%@PLMz@_X>L!8w2h-K29zcQmO3lH!abZLy~vzP*j-5W|L#9A_SimZW1eAO_l*bhIDUbd!O zYKWae7W)M$)K$5E{PYqsr-Grz#=Bo|685$=nHSIJ&-M_Ex)tG6)xD+~(7yVM;_g1! zh-o}WZ|c2}DeeoT0Y7U&^V!mtd+B&N5n_m&82TwFYco4KCW3~1k+?^JvcGlK?N&89 z!ZZ=>${`W?*&j%ZzT)c;^3NFrL7VzN&dY#&|o&VGQ^hLOE?P;iU<@ceZpo2FECag|{L zTzFlela{@@Kin!e?APW~5Ycl%eh-^gapv>EAkHAueV|u)p1y_yRyGnC-S5kW1h2i~ zc+x^05BL7eVm1P7GabC?k-S#VGxesh;?gTy&{XAAnugT6I}O^Co&J1P)R3%oqPK_W zKEgUkCr?Y&8(xr`*734~k(d%6UsMm2*ClokIMxU(UjX7fpG&yxHCU0Ql}Op+d_x$t z%DAGWiK@H~Zi;f05!fRr_*WQ}a`?v(UF1Gr3!zM1>S*rkd4wd= zux=2;vXC=_bzQsOA@V(ndwhPp?mPW7v5%|=GZ{t?ywXOM{RQ^{gx>l`AX_Tlsz6dJ zcl<3ob^Wr80K!N6#pcoNU2-(a9yA=;d1Szt@?aNu0mN`T2 zeS3)#!(@vLH%vFfT4emHz_%iy&!7@@o_H* zet%Nn+6Z65X1UrI#n{ZvIy>!4A3iTl$npi3IX2QqHJG{zup-lA?ZKy`5@iwlv9=RA zp>4PTdsJa0LhkGIbgUsB&{Nxn0mEjPgM7OmJBpg5v?gYv0$5QerIrp(uLng-&h|Jw zODGYLJME*fSCtCu*g7%3(1P1*Z~S4nVTmG-QyvYi$dFa8gR?N zB?I$tx_#ZD971M^aHPOO+6>cPil6h-NG3-*zaaS?|6Znr9*DL1eq!AmsZ({^TT{=` zu3?{I{enl@D2Qtxv@Ps#D-JZi2PO$bfCrj!nFUzSjkfr@)(-&Uh?tS5u5(i%up}eG zU%fXrk3i|(3eg?ukI3z9oa?LJU%jVm;783!;jXCk1?O(;rxviA?E3LgqmT!3@BlkP z7k*2D4lMA=O*@wa*vjWQ?Gayt2YDa?nYg4b3=#|DbPuex=@(wpmVXGi(>SKJXeh`| zC3*UW;9c4lfk>x-<;`~B;(og$Kt-q)0*NdjBLG7|b|?5GpWb+sgvM7qd-*#1v$cNo ztz-{`q4Gi<<6v^mO!s?!RKLqXoul-u_4%@*A3kArvO5CVlSR!v%E?6iEZ-hX1#3ju z>r{s?2I583p<)@<=yEZ@Gvw~vBc$Z_9Ki-ge-Qm31r0QM-%-r>YnoI+KX>O2+WPn5 zNL-{?P7&c#0zG&Y3NmJvl&eV-bQv@#1*YFycQeLR+F<@RpN)MnSw%3>r0}yLZQ!wIkNoS zkMq~?Xl+Y}%v$$Lz+lf$_~3kc`ym%y)2i&v;|8ikFxDKl>Znv$}JFm6;mebagb%6;;B!H^v(mUbiI zInFKr3xJ%7x1SBH2*gWH@hxjOjEWxe>3I%u@z8vczn{NTJ11=gqJ?HS^-b7|*+bsl zZ9hR&HUo{iw^gUMzxE?2gj1^QuQ7v)i{B@vJ|F-g2?-2U=!x8NIhrD1#bvx+Lnn1l z_gCd}=NQ_6;>cs6$@}7!Re0lQeXc~C^PP^%@-Mdt_gkl;U!K&URGnp{Os|9XLZ+ndFP4o^Mg zA)zh+PIRLzEm-6+8jF6V1w@?^eXqDw-lmloe_axC#Z^aW6@Um;4fcr zrm?tcwT{;a-Vt~rtokxP%Tb1w#2=9~HE5az&^0HM$>)B;Tf)zZ>ZHIrtn#w0`qt~G zedpZ{HeT@2`#TTLY&h%(YD+SKxV?moXCKkXI5gq${Eis3;}if~B#%^I)b9+M2{iuu z;x;qI4K|HH_(GHCs3OkGRv*>QB_5Qhiy4b>YclO%&kDZW{;nUA_$%-VJx=6;X{1T``!z48GyvC{Xjw z6Aej3+lO)}oShiEV361Lq#XgoArdkr$9Ug69P!;f#{>0~J~{|y93UR27?O6jA_{Kz z$p9~=ls{+LJLI$Oqh~SAPj8BH;5T}v#WfR_5NHFWPf3Sa7*UXuu0(V#}2?e7^|}}i@IaVlww~4u6YUi zJRe`u#TYTeqKRGsK}Xwvi2=`AZMl?xO&Q4^yE;IWP&r}VmV>(NOHtFpEem3SXz-$= zc5i{wqS-4)QcX$4I(ASx^Bh=RjpVC{FlpKr^W*`Y^Yry-xVYd{eNQ9xZa>U`2<4z& zr}_7gK4?<_`t(8>W#eglYzGcZyI{UN5(p&FmZT^j%=jdHSMXp`0%rnoC>ZFqWy`OG z?}k`8d|xp@uF*>iUFn?I>e>YIob3R(Y*RGZG6C!oXqhXKf>Meh8#In?aZ11|Fn_1b z$+rbrvM2qLSPaj<0S5*_L3^@8tMP6vwevNdPnq%{y5rkOKg6SMH%U{CZw(0@x=2j4 z>ruZwsW3>i91x#@crri)({uUwoSI=TD&)MA-sb1=2765``(R`cvyyho!#gMb%_eE4 zNW|I~R{`;a3Zs~Z)4POVJ>~DXS8uO~H6NxIBoh+ysL$Q(!D|qO`~)K5$amox5N6H= z(4$iwVOe{B!i_YYZ~+vZYumCrTje85Cufpyy~TRgtI98!>C?ISY7(7B^4xyn9)qT0 z^DU&6-x0rcF_7bU^u?!WHo#uq+9W~FQXjZ$SiJpme7QF?o8Gx&ZKVoDQKw~M>;eEk zJ(7L7zGEzX1g8dBeTpM#nAK+Eet(=u=yciXzj2!8S0dJ;%D^zaJ}T^&DxE24Q+9W1 z)-l{laVfj?#|y1`WhwsX9MJC)CGYdVUT`=iH6wDk#1(6xTB5VZD%-K7Wi+Slq_y9?szB`}bg3J+4e!mC zX{N?jMuQI;!-`%2Ba~Is=jp|UDgSD-?&WE$qUiC>y-YGp7hm-Hv~sjvm)AOI6G8b_ zMy#23xS@9X1=jn#B3s~O%kQrtDQM&SaHN7e4?i@r@X_7 z-JDn1lb29>oeeZY0r@b_y>P3;4#_%Woac7*)f!L_zgoY15N_8xS47yJ5sEQsW*8$2 z2tt?6eh^c8c%1fa=hGg}wDv0q+30)~eQ*qPaJN-fhlbJW;A9`ZOfIp}ki7eNtnnK7 zfpgdG#}0c_i&kT%(B~Qq_>}2c3GELmV?4fgxPz#n z?){_~4V8I5O$nn+#iF10yX1bhhD#sZjF8U55P`Sdb3C3g1Ssr~MY`?uK>y}EsjCX~M6-Ebec%!F400gMWgr? zpVY1r>Ucx4A}Ao!(s*@+ki<-}d$cc1_c$JVyev?&Ss$_+|*}VxXBFBwc||2G+HxRM5uoKvd6>M`Qiu<5=vNzszkWU`W;W!XK!iz1}OjHws`woGXU- z$bmCY3;JV!odjT_=oe$l{0IiGQjEV`?`}eOR?!dIQ&94WTYk*kZSiA5PsPEAoL6h5 zfV19f$f51){Z63RD z$PecF=Fz0G8xh!kUo-On)?qD)4;w{J)RO0Sw6pE#1rE{ z0r^sqv+7gP%-CSfq5Ki+>53?@h&PR&)A+qH%b`QPi7V8{Ydq;m3GV+ubA_ZO)m!uD zcEwrCdVl!|wu$=6G{}iHQxgqBg|g~1ey|Ucyov%5av{u@zR9kjk)`_md?)yQPO^7Q z(Xk9<<)b0ILFI_UC7p}!OyB4dSumJ~gxI|HDV9piJ~JVKx1?YquA0-yxE!(q9IfNc z765NwQeC7Fl>;0CsMO<+d=Ju;9#_?5Zq38av>dRUb$8qq&_ST3qZC(+i8Q%R6$70c z^aS)MhNd-^?JY%vu_+ASfH@sJAcOV3mPyK~AofYI26ksz_Q~3dG@Y~uK1+HR+;#_O zvQu!|*T1)d2Jw?)XaqIx*UW$nr0^@0LLmLNv*$dJI?4*o_|ll)5kRafq>vm3OT1>H zL&>T~=7RusXRFk>%ALMKb_(wt55vU+y&?>=fgj69l^3pyxTWoh6!zS>2q1<&IGD5b z5%PeWhE$$Y)8u=X1E7$-+(DOB{CugOpbbMS+zaF$J{XZst_J)i4f!enz`?h5y{s2d zXL!7wY{WMk{UaLF&&$ueJ`&%eIkO}ethn8s71gaf-Lc$q zx6-Hl@&+I7FB{4OTn_ImmbnDEzJNBEb8#Gce4B1U-Vl3@YaeF>iNT|m2WM%#7g6cl zMFbFh5$dBkL;2Ve&yo?TSIHof9&QMxz|APthZOAs;*nmN)ANJzI-nQg-@`U8plerqIi@?iHIF6m9eegb>YIIc1 z;9A;0OdWZ2pw5rkJq*#S50_mmhdxrJU*%m>i>w|R&_fk?<-GA*!zObV$yb8$JU(Jn zlvG)xz>~Z4lX^Z=c_bQyU@Y`8n-IL>9F1qt^>MlwpjN89Q*Gr3T;Xf?o7(V1m8fcn z>TNJTgE_pmd*3B`F2cLmR+(we(;)bgXuC=;hjqpFiXE!_x5_}Z9YOrN_;;)Ed2{qk;O?VqM>-?rMZ8Isd-|wIU}q| zN{@8cr<5Rn&XSTJ+t17EX@yPi`$uFSpKBxk!TVNRC=!_4PJjj}pMKmF-6H)m)}J=H zC~k4DLY!8zC4}sY39nJK)ie@^)&1iRoe>W20zQrF*+pa$5cBS=Y>Y>jYop$j+SC{| z&|CGPWjtCK{!ZZIbhrrDCiS!n07+HrUiJJAy-Ow%2YmHY=Aox*u8)S+G{Ev6-iP({4wV`O>&v|pH%I8W-VHgd4fkmRb?i*;-BnbVsCsd2yOV6($3=w$N!P+!)g zRl-Mde>D@Q&5g~b&Fkk}Tn{^#y}&W0%fnHH`xL#Yb>1gO1myz9-_04?ScsS79q#}JXgg>jHqu7h-;8Nr09Mlz zQ>`>XtfdiaBM&AmOv!T|XCy->sDiRzncG5iD%Hw-Nyg29u1?Pg7`({$Z`hVC#I1;nf z{5`XAl?3>pF;4R6C4B`3UTGJ3hTF~Y88 zk1J-XeFQ;}odR&KKO0G2rTYlyHwSQrC*b~(W`8T1ERp6D!ytqOb7vP&nhJh}?IiA> z$SaO0Su17e|1}ibD@-8UUa&Pl2&%2^=Po(}*SnSaP{3H&+gjvc0-eC-`_a>Sy3}!r z8dU1FZAOaiS7J0X%XO~glx?1v+-?19wTs{B@J8(}oe07S=J4D!cGy;Kth)$h=w5ZGp2AE(yU(VyvalntD z#>RcqXiZL<9n15}q>nh~zhgZNu7*f}IMmcnL=Oqu!BmoSmPv(PeCqGbxcy4~S;xT8 zE+={}&&m7Ir`i_wXt!K+819E7@f^h++-qCtzo2J44RFYw{3~6S?40 zMCd#-i2V7AC1<~|zp(Y<|H3c$E)?QFq|TlCUdvx*dE$E3q)IeT50QsNy}RE^D1)-S ztGe*uBC*ya?n@A|UA_orMeN`!T^xkXIf(c}s(2qg%=}vZ5`;-W7oI5i?*REiXRZE9 zIAVWpTN)~O>T->obuMDgvhU-a*K$bfGtJg#KRhPIkIAz)yuAxbE4PCo`NBO9$*-#BGNz~^?aIQQUA$cul9Zw_u&E68&=6FOIiS7IP17n$WpCO?#p8*m891xT~tI6)QS1E6u-N!n&{rAQ1%DP)H!mLuA z^a8;JnxcfbI_iZ?Sz-zg!&O@b$SFTBJ+JBm^yJ(R1U`mUO^pce;3h!g_PlgAStiXi z<5?^QQbxKbFSsp#Q=s#~gh{x3p+S1m7ZQ_HG2*VQFEvnTXtL*-4Mak6}dKbD;e zX94@E(zdMBc#uJlG92V6XD%!>w!P1Ga&m+(ADIC91ToZyL1r746y*n9jG{p01X1z+ z`Q8}K4-W5aJu01oFuR_!lT|>Zbey4}aZb;I+UP>tkiYB$Vry)JcpU1ICXW}cFc80M zFK6;t02+m~GZ_0IA7Tf9^wF6+#plny@-(FJqqS-;jFT!nzS39&PrAQhsh7&+=-63_ zJ}#l}OM&!dzhbAJVTY4^ukLw12JGxF@q6M11TyFKz8kq$kLkt(PHWi}lj2@6yY+eY zw2wA^SvnDWEjDUuQMdQQfa6EnrA#$|iD zTUWUW` zQVp?LusipZ)=qde`#QkjpeMIv8a-3|X>hYXlKpX}a;j1H6P(#4<`L)lqNz6bw+??t znCnp$-?uBmrTfpgkFOIVpT!Q4i^jrQ`7* z53DLK;*t&f=!ZapFYF!t44~?^{pl~j70PknlSu9xoXZ#~K7myC74rGGD6|tLou3Sq zWVDJ!WlD;a?D@pNb{GL_V`}RNebftiShq_Cl*30&N6we^F4` zO5+=)23#o;DzXaC)F0n%@5axHhC4(P>b%praf=YE2k2!aJi_lO`)pqG-ZG#Sf}j)^LQ)l=MLR4nJ&8HE#xYl{Kllc>`Jxm= zFYuln$z@-7h^{y1**-kR7cW1hhExOP$T06{y?}Xq5$W=#+F{c|ru`cm$W6VFY_QK> z7bH-@M~qiPcb8a$Daj*7os!6by3NxNv#d zoaVLO_x3u2dNx-UHR52kg_BJ0XO=}!t^kuIT|+oF>jcWRhZ4cVgM!n3p^Y1-@sK9Cg33zxG4{ux}IGDV~#qK+gO zvGHBC`%DXJJjgn(Dxgx1ujq8KBa#`z3sM+E$XFDw;`D0YoL;SezR~aYdfO0+(8xy1 zLo~;{bZ|#iLChQFe5h3vwQzi0eESm!JOuk&940uM+y^{r57LS_wiNLkw!{5QKE=^Q zB9s<@!9_wbviy8J*x4%ebwC`!@1j=5+yA9Tu6ceIOt0+u4J8Esv4}&h?lEoW0}^jk z7Flw(&NWy4)47>Y@53LF1Ibnu%=0PUU1!@)l>lx>eFtr))fS>X0>L{Ohm0u>vb$|@?-`Kgl^RhLO`w670+<>0NaJUo{R6`T z-wLKsT_g3oR`?nMl$bru^f5g@Cb@{bDm8Y2=vls`r%QRYX~iEUFGDDpM=&Gg3U%y^ z>*FIlaocKH;MT-7xM)>zxGd?{DQ(8eBoSGV&>m$ZC4I>GHy;-pM8{U!*k2rn@8flv z){JB~na}d*Ls5c6MV4Cs-9PLhTGa(nU>{$n zpP*X&ZUjyMCt|!JRJN$_4|BaR2;@EhEM~txpYsnb+c$z9RZJL(ns^-b3X9eC;f7hF zmp~kS<%K>EIK!r=!``R)NWHawk=dYoVl8g;bKE}2LV(!+nbc=6nb!T~^R98>D~=Pt zuFw4>Ei+DK+O7`-Jg(JSNg0g(B#}8=jEy6dNycI zA#dJ!WgDi>N(pJ{jWYmkSOGDS1Yj^ftwR(!YHYEx37}no47ITqfq$vPJ1K8;fIlET z)pH3O_1qhDt-e*P^qqa9o{D;sb^gE=>s3XW9Wd6X9c3l_6`FAZObfKQYxRmLEfFu4 z9hxHXi7jhya4e-@WKI@v(M!GGE|UdubtDEDVSTmXmK?t)!P|%nP|SNapG3FxDx??7 zTu}EQGwFn(%S9(aczKv0w62Z@h>B^S?|V5DUqMLH$Xj{&ERhZ@dN?=E$y2-dR)Qb?e=1Y%iH4xv*Hs2#pF(Z)eSMbOxrj$8>td#t(ro z91|qmuGe&ic>Xd5RT9{}ti;vcWe=JQGY}-?Lc^bDw?6Uj2ZaP=;=F2B4eh>*cozyF zDI(kVBT3))Z#*uc(BcIFbe|88yl7$1M-BSD@WHe=NsU05u}*6Fq&3}T>Mun|Hyyto zs2ARAicI2gh!KQ9(cokMeQq|iu?AA+x9a%&Hs8bH%${N)6~TT~0T!)+G9 zYwP4udx1tO_Fx9wh-33U<2CJ(pBx4_BBVbc9`~E&`xGTPKPB!4V8`n9(XH$8CErUs z0GrGO2E%OdaI;7DIAPh4mGh881(!1W0nF1r2+=%`DC_r=09w*#63MuQ4f@}piFscC z_*K8fBu7q8v!Al0n}=QrPC)1-cfR|I!~HLXRPZk(Kx@%ce>C!P46k(_hynT_$Ulp_ zTP%KDK-2Pqv-JR@lZQV4ED{6x@q!>EJ;L=K^-$E}!8-c<4iWvLl$u>tWD4=keKeov z(;+$0Ig1Y0-f}8jVc&@H>y13**BaJd*RJ^EJ?fww0SWYe8_qnj{ebA%cUh-LI$^g_ z7el-oW1woKz@bO9-pg)!dhzeejpGQ-QOJGm{VM3`F6*>iJhz-~5guT)e2-+Jk3Ocr5#A`Z3Jp z`mDMv)e{vJL54Gj!vN~hU(&S8Pg3?ee0?9e3P-vArWaX`eaJZVTLwjs_l&T;bh2{? zzI&_(!wfY3w~Hd*5~CL{{c7dP^ne^wg6O~Pz8HZcxcIWZ528lt{pgaeC*mnnNYUyb=*-^{ zJuIo@ObwXoU~#d_Qwg}!vkLcqz<%?T2Jr1FTz8kxG+GZPa}dIz0LTwY*~+c1(!Jf?O=#RIlK9CENifC$mw2ZgSA9adB=93=)%Vquz7b%w-@0;P; z8;=A!{ITVF>T)}l7_a79&_knxS3-}7l!oNE&=9xOS8~S2>%Ios%n6ATz)c>!PYaZa zi1b~%GLPYYjkS++@cs}t6s^;AD3TuWM^`D_7jG|qMuG{9xQmjW2clEukaozN35KWH zFF@0dlJx9Wd-WiH_Tw+hLudNAjx^Dw{x+Km;zq<#(Krg8fS+-*;DG)Wx+G4P(}k+t z2CtJC|M{-w`h*G_ucym;ge-V|vbP6@{!m)}oxv%90~V!RGiPbP_A#v?4U)>exEi0c z5KQYMV`hRpxNQ?@Unbe+hC-+rzK+4CS^|ds<-L(z(}P9B5p+c67N{37tcuaNFL!^x zB+v-D1R{8Z&6++Izz{}Qa?6HxTN`vJ60q~$+z%-bUXTaB*7HuE`z_6o=#V|O&KuoN zSVu9b1;;0DDGid~8aO>^e$QP2_+kBg*cT5m_g~6dkOApbt1~>>_1LM+l&#S8pBjX1 z{b2e{|CSX*P*z*r`wo0pI>;F>vW6Qy^Jiz8bQQ9)`*e^ zT_2Q@=P$Y6F{iIA8^$O-(f79AJu)Nh1+yLZTo+q`@!Q1(qH-vVW1wgOd_n;;IWBfG z&bv1BNf6mW4P`s0cEBx~m6W!^4P?-z${?xHN@v5IBC~-4H%EqFHaF@e+ZXYn#hk;b z9^x%VcmLM&*#5%$byfQ?Pb9Df_w(%uiofIh!NH_zU}nYN^b-R*Jsb^3qAt!O(%+tG z4Jax7UEj*0BprWlL%&tgNxaP|)xYcFn}~-b!cf;J>6#OzRD3i#OFOE@KLXMb51M@r zl!J`r5_+`k!ujVZVAu*vZogyI=KBT2o<+53;{A|?R$RNdWes!VXwvKPx1wv4L zFGG0Ggz3aNF0k$_gegJzCVrmi`Fo{AqQ8rd_D?{?w>reoS*gAHXZC;>t=Mi7-cNxOeNBWSS ztOeFQ2kN7 zh0|-E)AN1migIMO>s}lkh8g{6Bc8(V&|L($k`siY3;mzgCD(q)hC^CoDWrA?4MwV| z=ozHVeeWX4eI(@7j8_R_9pHla4WFH^SjFkR6hKI}ZHPg(?08k{_4HB@)^1JfSe>cQ6$pSHnIjd1)uFR#}*W-ZWI zU-^X52;UAGJcTkRoY0;_<$hnI4=E0UskQzVVIs)A_9+KPuaA1?T3;3Rkf2+p zPQ)90xXkDtuCYb$4sk%|ORD>mZFvF#A>xJRTb-YC{G@4@MY(;_{VEQxH`V%yCJQtY zzOpB2Vum4bT8*-;KT;1WZ7sO6m$H5-4-jzC;c8R2SqVaIYEmCWWMFMojmF|np7wGT zyBtc27^f-s2Dl0o=ae%N5!=>>gTPV}R`kYQ4EW!17 z*ccg&;v76!dVEf77ZaF5tI8Mb445gXX+P`ef6=8WGDrn9>~vXXU}umtPe0K#)^j;b|tnGzxVo|7?jmQM~)3%;Yt$brda zr5>)abpY^yLSp?8t(NX3qhb=a_SkPC_BJs02+47L@gd5eb_J^t_w@c1rReq#xBuYL z=D*#0=hkzQ$|p2s_HYtKTur|M#|^1e1J_BwoJRg`p1LhQ#?Rs!LOlRMB~O}0*UC%xFx_bIK7NeL2zi#<|B;e z@6X{2GaosObJ4vZIS}oB>p|p)R$*1Xhv848wZM;l>V!fQjsBb<7OooHGXk?-v+m(0 z%`6g_35DvKdVXFI;QE0%--#n%;^rYY0C4m9p;vah1_ z{zUsGVQ!}u3g4xZ@9ept*-ZrqnR5md{^SAymJ7Y53%RX*tM--U`sv6Cw@Rn)EqQ>5 zSoGddOyZ7s>(=*~QF26#f-X%@hdCth4y_jL0p%AJJP0le2@~KwzL;%dk&7A_C^}K! z1S}O#m(Wt;^m!$(>zqgUy!3oK0uM#3ezR4i=5^tj>PRO2YbDT5^%Z|kra?t>G8;S$ zgf`SSqmEN~pA8?f*(k)1$p@S+9`swQVI4N!HFN~JOU!(RX!F|L zT7ZO|uNyJ_ft~HW$wFgZ!+Ai2_YTgN63d}+507^l44ydBJ$QfEI0Ca2pCichck=i4 z&>7$!DLwn64KO33yTebV;wy^opyyR7g%+l-?KQP{up6Bs!p8{BE9)r2t7L3HWTRzQDQj zE0Z4yQwh&&GQRwee|deV!;J}#0BF8YC$uNmv#tH70?6!4V?32C3yXSjFJguG1x?;c z+)V-@T9Q)w^gZjphvmp-;VG>PjD3=!e?0nnrZg>oTbpzBcvrwOl7BkpB zvJNf`bdA!1xSx`X18CmwrY+~?`>Jxm<3FmS5!Hy$9HdYB!HYUt#}d*6#F{(#z7xP6 z*%H!S1u{*EbBLys`=Tj1`Zy086uOx>Lenm!*ktg{=&hhT*&6fJv}=e0M9qDBB2bH% z9WM(GN}f$2BN2=#2%~hp$I|B;lM5Ef<=0-Um%M%z2N1+cSMX@$n<}2SXmK51&PHWFVNsC>BkMSKrw_z+owFG3hAC?>Z$Fa{3LY@E+@{ z`Ye=2J1u98>>h{N-wWa;!R9Z)Ggv>0E?AQ?I-Ji3DDiM9pp+un>Txo`9C=y$B;t=q z&K~x?SXNQwj4NyN?uwK+2>9@q2N_+m{Ps`{`O+(+d=N9Lz zHJmFqX;99(CIIIb%)g{Q;GZ5SG$RnZ_DF8Z*~stqOUzy}5@6qusz9I|2D%H5%=~UT zs1s1Cf*pentzHw~Et9Vq>CoZa()`0|Z|;*Hzs*)0)t)Bz4&pp_)U zXaNAibO>JUiK=~$ogjFr;*;eRNQoP-EbRAC1m5wkBXQ{|I52XeD)U`TxZy0T0)0l- z9AK)e{K&WJ!v$hFjF6g77zOs&SF4}*N$#!Az&4UIJ2Ql@jW%yA#p$pm0C#GuYJM+00z7Ek18u z@L4T{H;{5>Kmv53u8#f}DQ5csr8}bhua9%365TDg)=>!QPfd8(j!Qv~0e6rsbZd-A zC~6tuV*BE~fL|NFkrg_FI7Ns|ZCZyGG5z77)pN?fhOMY~6r zm>rNWEf~qx8)?sb_S>|Au0+{$Ev^=qOZptJt$Yoq3u-XqD0$HWLms$?$|J9eF0ZGY z>ALPjbx|bzm?x>}&L=G&`^#*r=I%t|`*e&N4tk7tj>}dG$_V$=szRmW1Re*!Kz$?f z({;aSc=W{Z)BT`#ua7(SNV>qa5#4}HmEN<*q25^c56}bW-jJZth9o z;+xxB3kBWF70)!|D>5#B5nc#*ZXvir$r#V?71MU;qPO=h>Vr={Ul!i(_s4xxiWKe7 zS`T2$I5q9V+^FAFEfc?@iaYkXPaz8H%9VLf8=-+WVQ$hH=jY`V0y~tE>u?H`&5a+o z66;G1NdWi{u9FpZRt-^t_Lx*D~uF72h+&sbwt^&8j z*qf8MVuGfiXSMyPrjjQnuq+6Uz=<2kH?m9qY!Y>Qua%gF!G8l5!e-y?Z9@MMT}_0z zQy_Pn{8znZ$0~b#U_XVK#-Ad9Ht!B%crC_8;sdm2o=85QcYiwXt05&JWFjy1tJ(F< zwc}-wUbGuB(n$teTSxOS04&c;la@ExfC;t`>H4=OWXG>fEtAq=j5FCPv^E(eao-FnC7Ixx9a0dr;YN64}iR?B>1Aty~X5It?790n7GHJI=@UX z#{kh;9kE;%`vFgyAsVhYdcSnP03*N`=(8BDeC%-fi(ZJ@1wr5#AXK<3MQ;pc^@Lrf z_7u||wy&N{=|sv``qlT68xjXsL%d!t^1u7{lt0~^2|F)M4yZt4QR=-$N}HZ)fJl#? z-cV#M%*((L+Wv)wL(cNLw4PnGskW0KD0lvMaM|}T-4P6h-xIMChIo+#G3*V7h0XHE zej($ASkc^mZkJ2|cCe!D=UBASxyL9vVDMEZkxy-N#g^Xhg)SIpNzz!}PvU7%MX$2I ziFW-LLHEhz`Th>egYpHDByDA_t~;Nflw>^k6hf2&AolY%&l^{q?|;0?&GWeZIvcw6 z4!9h6RCkE%t#+R|9jR3ra$^~DPrV1~yZO!^tS5=jeA*5Xs9Yr}$e!$6EgUv=K-t9U zjQW86$L>4ZLuVVgn9Y$_G`nTZpf4dA8IE+k*KNa*IL&BP zR+(f|3NSvf`h20k#`_Iq9)mP-=mCohq4^lT&mIET<#~p|;RvH?Z@mvlx;6PoeUT>X zJ4VP#BG49YGapPhLmhOO_=sCFk>|$m{!fZDZEE1l`LwmD_cevcLUO53qFs z765Px!k)O`6ms5X^gM`qBDPf?oQqNn{t52|{I&6GH?vdobCBxn3|svI;1< z?~5u%66p`K{GZhx|a{mVDp@n2KyOjM^EZRU)Pp+>_fA48}C(x5?&B0c$(?tWaZv|q{KvW<78AxL5=(o zZJykGe3~G69>gE23&1kHSs0K;!WIHoa*huK!~S^L1&Bu%=jrq3i-`%8D&+%npv*Pe zUCF$!+BUAU0&dW~eO8|UTrR}*OF(KWO8Uczs>Ap9mgNLgfv>(H|Dde3zPo>^`Fy5- zEBJNVrYS;tA}>!;zwHuT+nW*5)9Q+=OQ;2o>}fi_ghHLN4OlT&))8LN7U0LZzLr;# zYrh$6f-%*#hX3fHl}1hJ^gQm0`8Pc;m=4nHj`86SuhbVI4?A2OKt+W?sk14l{Rlds_n!C&-d z{sbWs-|i{k!;b2!@Jk}16l^3z|=NbE)TW}4ZZd6-J!n}DNZ-Ja7N)Hk>2L+9{5zhOLu!Tzf(#tWZ zH{rdSm|wbh^D&U6bb)vDxJyAKWA95!s9&^Ky6Iw=l+G5cd=&TDo+HW~o@8Q)DT7Uy z^JVpCyfG;&z4khEAqa+u`NowvU65`+()aGpkpRF`@HZTv$=to(JMpI|f9;H^?fp>V)V9_XC(P z`9)V(V=peYd--^o_V*Cnq!o$@Kjz_zB&>#2UKMtdMZGV7h$8V!#u|y#itL9G014~* zqvL5I^*ZSO_?w6+UqQ&VL8nZX0eV-SJOSe#7@N!Drt^P${TJplV$8J~*x}8@4aUQ0 zo2~?fr!3dGEjHLNq@p_;B?<LV>=$hiLO3iO0Q{A zJ@Z2-D;CraE(>!{eUk;w3O|aKprCGav6v4y|QRy~BDz=7U^13Kw6${I>Bw`+`(U}kDO zHMlSk{aRFhubR#{xTevHX8C5mq#tSHEa*C3b`T*$H341mHM$9s>X0f^Z)J>424ku5 zn=AYB2BZ1~(n=-1k228u4?Ygs=f_*_6nai!9b={BsMlz|)jI5v-K_vR)p<|;Xv4Pq z;yR2!`0JsKRY8r1PC>eg)*#%Yjpxwhkng0+ofK!>C0%GZ&|&!l!bCuJDjFNoVc&oC z#Y-jt#HXco@E{EJ!!1U!=Xyn7B;%3%=ZV=5qgX1?UnXfe*_6`@r zUgG?sI!<%XEfl8IUP9e@x_&4`MYc?Hf z;R(_xpc1!SXRqK7^UX@O25cK2sa$`xwq1ah#lnq!>PrGboR_P&!HrXn-{-Y(%e|0T zPpPJ4Oh5xOQdDgBG(qgLF?&25FUcDl5c|hWfnYE1l)9__$Wc950$1Vu`2p*(m_|>B znf7}i^^5PR_LYizPftC4cxFGbDBY*{02`~qSaH`8TVNBPuNb)RAmtnKn?s+r_i2#e z{L4P^*$P$o89}Fn{u^iHpIql@b(`X+kHmGD$FFp`GlEgRY(SddW;}yA-Nwgv-xx49 z*BhiO;YjPN=daxjzx(<5z5({--d5Ww5P{}rM@TSdIvOtDMwzcMb0PUAx!%)znX23Kwl8f&V+|Z=khi-c;~nD;4JjHg|s@ zedLT5aX`wNuCw1kQb*)IF_EHq(k^=w!MO;p2~a)^#eIOkd3MW=V0Z$3e~EeriRR#l zElJ@@SO4}#?4|T;k?>jGtWBl%L&q31Z^XI;v|p;r;=_ri{p722v`qWL&F%}_?eN#4PUXD)~aO=iHj%459G+%%A_ z5a0s`85ixH3QHY>AbI&Xsp2rk*;)2HVQ6Wg(f`ArdGHNVP*>yw?>okDHCm(uvVr_X z6ye8v1)l{<_C{?a>cjtFt_(@y)XaycY#_~ZPn3mPLVaI`ycgV=xNQXP9!tVKr|2UX z$avDUOW*o>v>*Pg=2cjRKf~c!5jotQ%U0%kMA9uiPwkoBJE=*Cq3iN{e+=q)rgSm@ zzgs;ZzRHi<@CFdg&7bMuUh-Y^VN(;`JR9aC;hy1i6lNB3J=(9|D)!jq+UIde0aheE zUxzN0x#$P=&<%8Qz6(o`Qxe>-{&yw!TmKD(%Ax2c zn0}ExCK5k!ynel}d)dDTtO=?JmVm3f$j=K_xD}rPRxXah-1fJN<7>D1KJ%%+K?_xI z=mr2l#iHXFo|40o90u`DoJ8}!0j>Rw+VkH3B8%*f>eqUOsbtY>iljh93{w>1`#`rk zy;Fwn&MuGJMEbdJc;G&wj^c#db62(Z$3FSZ#f|T0Tm9uZL?zKV><7bf)^Lz> zfmq|SUvN|fIX8Qh5bN~<84CXfzwWDBlfMz04V+t%7C5#m_$54LtQ%ik^qytsJ0CPL z%791UkD(=U9s@(el^WmW2kB%BYL5jj1#P5J@Uh%C`er(rL`t^rcs*SqGQw@Nszw@8 ztLRXmBN2z3UI>*yq#1v+g+%&Z9%GUe zIHWw>c+V|tNAxDztKWeFP5Zi+1N7eBIfF-FZ~}eEEzw`_hIS3YIU-bDQiX?pctwiP z3>P5zywI2f<80PYJPZIp^13~UC7W%`4YA{^u7$9C=t9eEA7v>_x(g_|nJmkbAJ+Bv z*}8CXugGjnnTxDG+0%nwF!W<1fA<^TEMa?JF3ID=T~zsNdWXaH{udq9Xdg`AFo@LW z`$*~)f*f@mA0KoTT(x_dgX{)>?~=T?I8)2w^l=F!NzV-wi2DY44KD^yQ{qozZK<=) z{&*-SuMR&^tCs~bv)}P8djIT!#!AzktM6WKUiJ#`c69+hdsS_!b-fXv-zjps8_Y^R zU}O6B&K$4p;eLRGL?rvm4;eQQGj1BjS8~B}WX>k62~9#c0F0ee!wQSf+ez&wf`cH& z3f*Y?!wzV5dR@_^=@G8`uw!m2E=Y__q9ctiDefsPHfaUPAi>{;t2^NRAf<4_*&VNx z^4YI>;_z{Dk^Y6t8{Et*@12xd$9xAA22Xf3>|uB`Oa>n_uYT~I;MMy?O1b4*))7!* zII(YIpBU_g3ig)x=bW#R2#qWn<)c9w6HYRp=5;387ndi&_AcF{v#z+W@L=zJ8&sGM zkPgjU;)_gnSyCdlLh}xVjTFKIqo>3zFV0t}JyztMBEi8yPDQs9N~DrP_W2Czkro_c zVn9Be0HdCQ12huBpR}`*p82wFnm0ikJFUg_YJV(=4{+|n!7Tmx`w`*d6(H1~F2=$o=ybgiVUXNlT8goxaIQAS)auWB@CZT?oNce) zDymTWQ+nj>2=Dy zSl2kMuq0n&#~tfUgzAYYQk-s2q-=bn^>KURY2oFU}}$VdpUsvq>nPPzxKCpza$=SN}w5* zM$%a)H}U!VEYtP&Jk$FJCs_ot=Qreom_G(WP+9JEW2~AQ7yi>Yb91vZg&>L>&7h^0(bt(Nb+7Z$_k1D%3M*Mt2aFn-LG_se3wgl`? zZQ!rW;qm3OONbn=GvL6#s%G@X?hM{9;izOkHyKKkNyZ9nUZVZ!kuKJb9}Rjt;K0 zi27P|4wN(7uykp(9Z4v`@Z(7^>`42yf`2t+IKe4!(8>lI0J#7O>)LhS%NhMa0VO<{ zIAw?90CzZxL#9Gg4n!z|s=^Yz)iJfOD4V2TW5m5ck}Db*iontwIq<3_pYpQDO#Xlx zsk$qTSei|c3aku-Al#~A9M2?kuFlO@-w!!7UFP=LDhi=OZ|lx^TNia6L`bJqOcr=W z#DRN#&tXRJRCt7Zal+7Q}Fcw*=mgfL=2TnNgfeJ9cmK0kdjA9@_u+Z(`C~1N-bHA z4_O5t*igc=xh>|@rrw*279wv;2E2;Y z6Ey<47-72~x(%8ddEr@+FOEcRiYdK-)ryL;E0!^q^g=kI&xGK-ApVY;O^`7;f>1u_-_ab?}NFGLj+U z?6J0957l=r&+>8pE=ylIB1X3f=e`-lyC9`zn0%~h-4J-C5jTfF9OQR@hRj*n5TQdh z?^n8XOSO;Ff1UuUY96V)G>bSpM=W|s>q>=@2Ec~^++4iWSD>##1VxT79Vx?zW%b;!yya6QR^?CrIvhuPgpC zD}`^2UUPvss}9uE@zM6I4)M}ou*m+yv;MVEY8s*`P+w2Goz(l1W>kd4%B>nYZ)kGV zBnD-t>@ypC7xcsw4lwNHV_(%efyV!OH1qMVql%O7KbFpGOI0ljqQ8U+ijU-+GjE7W zmLLlF^|$DA$2jAL?qkNi*Iuh?mXfQFn>9pH8^U`rv|T|f9U_89it4pDq{@*1Ruz+@eICF$eyK}TRVV+mt#>;co1 zr>Gz>&X%I#^Q?dStLf@fA_J!ixuGJe+6G25I_m(3<9FzQ9w{pWS`t~UO}V_=Xbi#XDq-OX{o({ z`Wg*BsB}T%)@H5AD-#(Xh^+p*fQvaO%JRP9BQZCB9`U8mI|W$?!IdluLWggoIDF5a zf&M!c|3fR5DF#q3OxI}1DH6inJ17%fB+7pgO2y+|QIc;-Pst{Gu^!gbP?sx}SNc(^}Fk*!_wyq>PU@ob?K68HOh zCUWiP2U|BxDRqHRH=j~>qOYb&13!dL3oE1O>VGN|huF&wcmVW@*&BJn!cFh?a#|~u zaE&W(INo^KXvVU~j&zMoJmfm7j|hJq!abRf6E**GT=<{j^$tJM*Ck(k+H`=%Hm)w0hzzw*M@yn%W)a)5Wf`TYy z1b-mU@L?yzrLtKFC_~!3Xo9Xe@b8`a6Am($nWX638tm1-zV>-(auz7`L9Zbd+k}}1 z4`2k~L$!_nu*bFDlH6k9d}qIL_xMmL_YLTg(7Nxi*8-OJyb zM<%jl)r+Mrm@qahQ5()>WWKi*aL$9*J!=H)0{jn8@9kah7jJB%un3!{P#%v)ixt4@ z4?;|~UD75|{61d~r|ENfu;+(_V`uf=2mLL4IG%IRs}96Z@q;yyA3)w3#>F}XFF{%5=n|;QR(&u!v*)-k6Kaj?BpZ86QQIikCk%lmPO95bP zZ^?aV+P%tV2a@`~vJD0H6!s_KaR3nQ4o93k|4vtFG{6i(-h~@j;8}@9yS$}VK3orL z{0(0p?*LJ7piKSmpm}qtUt^?i^QVGwJAJgsT(yYeDvzA7SA2l( z`-FB74lv2>zma{B&RJ6Ae@w0x2C@T>MC4hW(LLdMUe8SadBzLUtpsvnpztZ*W9ffN zdcH#x7}K$>3wUJK*xp}QMfeU_?Z=ssAc%!pnc@y$-yigIA+wXKO2vB6)yOEAq;+}b z8>2Zl1v1a;&IwmDdzYIR^T@!zN7795M8ca_y1cxK*DTu3!NeT!t{Rl>BGlcYjaShzy2fp0PPQlnS&QvZLSNtC(oRArAp0T)!-w4SY zs|a3RsyI#$MN{Pa8Kn{j=7#S^yzkya2*wfXB!DRBY~ME$b({J_hjdCy|IU(WH{-Px zmeA=vC_IfBo>Yw+tENEuxbq?;C{$4aeGsHk0fU1)YbzLmPs@OU#Jx|_(R(aK53ZDr zeT`|ou%!cquR2Mi_`Ef7pNvB*e>8&25C(QPentz9%a60O**u^fr zGd21k)K;!`0IIuD(f%+p!7d)+=%z?ektk?N*S+q22X{Gg5g z6>zK=bftW9%{{`{U!Go4hhjm#-p?0mjc<=KYKiO@(^)OYqeUn@jTpUsI1}}Vnap+}pyoC}!+29pd0*>R`XwJHE>9tP=hYVY-h%J;n zoS*R5g|sqd^^Oa3Rjes12xwBVDTtOS_}LvWA6u_h0O3DF8lIf4P*tlSRKU!C3JsmHX1R}Ke2TA4iLA0V}cQK;f@K7MOm4?@`l zn@+sn@l+likKop?6QtHJ^5dPALW1I}n;rHQ@=&zr{ko9~I7g=e?=kNatw5%T1LhQD zDOgIP5_O52VZ)*p2@=(O89%@ba9QIJmsG}=#6NtozJI} z6oFU@d!I(ai?|X_u%OcH8z9xKy;1%gri{8Vg%cJn;%sm1kq4dYQQRC1^bnmE$^5U( z7*G?kLsk#uji$Y!L|nPnZzXbYb~qht`;Sii^6agjQ_3F7QaEuA@SY&?H9`ZX?s_s4Dis2rGRm{S$#;3*L!qFOZuCadp?RD zy^vpTAp2%9**8#W3i04yeGfxVAF6Pc?Oy^&+7|tgil!1Mp*R8qD~PZk@4q2{wf3NDVs*b?^$WKQ)l1%ch@*w&Vo|ncBdr`VUe#OP z_i?7Idr3>bXYJQ4fO-l91m`aUyZ_l%SMGE?fuG3(;4)#dE| z5nT=mJT^oT)D>julI#u|G{C?>%>f&waOW(!Nt_rsM}-Y&s=eyLp9KL_xfmh}&@*1~ z^7tyypX8DAO7gjjk|2`cps*nB|1o0yM1KedAsLDudcwj5IiT?1k`yG+bMtYLAOco# zB6BMRj=4C3Ylr3$X+FAXzC{Z+k)`F4F5C5!@&iY3wnY%><57N0-@1Q_!BoI^*&~ZJ zUng)r>PPn@r^1dR{ggb~{+ygYKdld)_2xlXS$7u1Hy3S3R2=c)I-PWG+Y zKQsgjjDn9`ua(wPU86nJvl1(KC?akqv_J#0Zf+s&v|D8#InRm*=>1S=Z-DtKC>O8V zNnDPtAM-si1LsCXArp>a=wWD$3kveXUrf&AO)7G~TuHZ@T*DCP_Lk3k=^6Pd|3Yw=NG9 zz&r*JCNFX=mMR}`XW}D(VjgY?YSx?ZxSxn=X81|9Lkt%xaD(G;JDzlj!Y}wY=ONbx ziU(ZONQwXiqmfc%($x~O%+l5MoHvse`d+In{OoZ3hx}GwYdL5b(4GZ=Qp|%6Z6^0g zlcPEiw(|rY`=R^+clOd)cg5fLT-g$`bG%9vbn45Tjpg|`swp16 zYF_IIA@_@fQnayNbAsy25!dKg@HwK}Ra%2;AueOaH|{U4>j{@8BCS2zO2okKaQMD0 zUtdemIXFb!_WPSCP_6JqEJ-o_fIMeB*M$TbPo*QLsCmoP7GfP?3IharaIf4pj>>QoMAQ$yy#N6SP9^e=PQ9^sbjA{AwcJvxmKuHJf~xi z(m}E4_db)=1E_s(H?*FLUoB2or=-{IN)n@N{>TmnnP=W44;eiS^|>FOes;jYpAULL zsvyscIsH_$7+&?|#$m-_?etwzK{G=V69T7oQp(qvd22PM#VzC+M*U$Rm|xQ0+uM85 zFZx#vDfnLGiwj$fC_rko$smEtH;gyWMs=9b=0O4WNGL|wV9Ak6RU zv|p=dw1=D0=m|vo;gwVuzQ!ud20(h6K8cP1+>#N$_$OgFdxc(YTkC2c$=t_KO#i#c z;h3#Gj;innB*IJUKTifZ;p4T;nK?y!kb|v-bGRt*Zo$ZtfCS){SWDI9)Q27YLY|YmiwX{ifb+8k4s`?55dsa^gm#WhKO4?R$ym)F zEi8hLz+!klB!E1*D+8Y4zU-;DkQZO%RXV*s>d)p=e;z#j*r3P2#sF(G=e@+h+#g~V z{K!l##}gb-+c6(#t+C=09PLg>0Wstv7N}D}Zfmoh5vJtw8pgm5s4vGaRkh0!;UZy0 z3*(~z4r+h^aK{VC2A`Y#0XriW7dn;>!3*6oN`=TK&hO~3e*G1nkQH@b+#2ydF>@Et zJi8OG%cI+mn?+6PwnU!Jr-}>Fd*_b)0P`f~lj3f{V-;E7i$C5LQKaD1aqk6TU=YFB zikp-ee^$KXLPH1&znLL83LIep!u%qJo}O4`AR+isFb@fIF2)p3KMYjp)03{YL@2dE zFpyd(a$ti{u@G7@2a47H94-O+BUZ6_2OPmZJ)=24n2Lw% zi&)%()+!xKXVBlbTVvU)oN;R(;*+8pqkY<`g^ z?mdoLbqEz<3ONP>KE>_&m^f1M?E*~}>_*yPh>r3-oyT-uU@waf>i2*Y5_KB_oPg!S z;b_mtED3@s7cHE+)TAM@c1Asd4V2X+`dDFY>lrTWVP&R9CY&$I@ctcnW7%T&WNI4x zr*MhKT|5qjt5M>Ix~KY)wx7*W)=Hm=HUPL7@Iao^_iRBP(X-opMSng^d%<9W7ljU# zeZl+3nDCStB0Sr()(T7q>duRL`Tlr^G1%;jwG6mYdAy)eqOWgUOWj8y6_hqiM+m|-b>S8 zwvhSLpZ`Ttd=i=OS;lY?1ql!l@r->GA#8aIuTosnj?o%Zy^1=Z`%v!%*vp{I>)D&OCnf9 zz!6O9cFf@*c&r4t;ZU^K?W4}tkFkheZfYP#z-jc-8zclI;tFt?zs;XR2A%VZ9JxNA z)(5Y)srIJ*uTXHXe$_9x%0B$QzU?oD$H_S%UQk5~#zU1f4^}W=buI>tBl4zoQmf!YxX@BeP8NVZSHX#!gDrA;Q&|l{LK^gy;e&xsq8vM{j`6 zGy+g;i>kEdBNEGIP2S`Nx5G5&cT z(s*@%TX@6;Ft+=-e>(Y6RbzI%X)ehwkB~GVy0Thi3$%0>zJ5Kw6Q7l(V?r`bGl$IpjUSc|HaGNw94a*au?Fai81k;uM;ApCbwGfe1Nfi-xY9UkWcZ}gVRWfGEYvwd_=Cikb3WP zz2aW=25w9sRwk| z0z9;MXgoh>$K`L_Ogz;Lqy?}k=tp( ze=gdSoZ$rsLau~U1-z-T+N%x`xJ$A@QzIkIH7<>+6`rbCl!=N?F6rSbd7ttz)E;)b zeiI5eefBeLd!x!AaypFe_cEBN&j~1#M9BPR02(NTfje$KQhBMi5(W=U+T$rDMI)Ol1gY4qlYghd=-nj>8(f*KKG)xjKH*K>w8dX{O8tvQD zXZpYC8pY@tA%fz)g~bripAve8PD-QJyndg`6EBV+=zPy!musy3;A^E)z%a26qQUR_ zy<$NSk!KH3(2sBUV-9NA&Vx&aQP%eh;=#YA+!xCdy-lxwZzbH2kFK@j!2J}d>BL0$ zaNFnC`o)sUjgmNA`V?-W)9|gLynsDuhQM4K+U#IBZMv}&lpvTlio3<_`%SzI2|7Z0 z^JR9!C954QVS%?Ld1syqJjz?ormqJARpO&gzCh1+#ciM8^)Vui;_W*yq^oqrK#bTe za2u~61~l_~-*jP3$`>&Cj)6qpaHH&}4nNwC*;3GCDb}~QB!TI)u*Y%9$5x`U!yVpJ zNh5&N5~h~)Hm20}((NtAhMP;A0e<9F>l zx?Gu2T@WzmV>Y5#7f=gD9I>xogoZuB<)mxN0Y0G~K0SK4TOnZWSKh(#jJ0^Iw?U_j zim`8Wa3)9GO#mXDe^-PCYHFx_DTTAtT2{?&FxN29wR z4Us>kl!;;m&TI$6md5uH^8ifg$h`-+Z7MiH)BWq4eGmt7)gH@}mqM8Cv!uYWk)^#} z1O5SLB?MRb5VO2qK81RtVVUF5%XUwN+($GZyNO6>o9l%EIx{m0U5L-_*CXg&irY8(9Bl9X8pDsf54SPX^n3OgrD5=qu;c1Hxy$3~-;uJLv=Oj5(2b&xUDeu;x+JJZX5(~xP_Fp#FrLba*~tHJ)w zTx5qvw*_)ZjF9Z=e0N^TyS&d{{JC2L?L*o0Xb`m$BfqTePe999!SzUnLlED;g6Y2~ ze!ZcQ5nFmGDw-&gP={O(^F<<|%YyfqxB=|Dp$E(U+`W+3nje~Ij>to;hl}paw?8y6 z@e50Kf(7&O^<;@Z z0+GZfK?9cXcRLj)GMoq2x5$oa5#9~Kogh6{t1pq|InJ4r$1 z5y9WiPFLcIu%=+<{SyM2kvJXXH)6gZOB1Y>)tm2d$wp!D`+-$#()p{e0*UF*7EU7I zmGZqz8z2zwZ?K45gfw`L{V_Jypt)1o3|3G1Es8eRqyY>GLsoj``gFh22ha&TC)WRf zuAvj$OGzP;{X-xS#YfuXLPi}+tq zsTtKa=I}LdmV5`wAPwF+tYx3iQN=;k^MndwmMP95-Fw$5 z0dBIq;DW^SU^NQ>7$}8W(uYQ8!D+&Y1&d4&Y}1v@YVIc+a-VpaiBjqDk{JZd&Vi-0 zesw`hE;xf=6L@7k4{yzGC~+2y-zo%A>4AWBx*7fye!g{72g3VhRotu>mx^YZ#}>@bWo!J9drc|CyrE#Gq7Hg9?fomGa|*mzW2MI zo$S#F32(!}sd1oW^mF4@tEwYxncY{4GmBENxp zf$9Jkp|~MV+AE9;af_p?_SBBXlk>jl%EQ|TDSk!qfXm^Touxzu@-t^v;@U6BXma4? zfGq8rLr*WHa$3>eG1{e> zpE%jUE8BZ9b#oB#B`*E;fT zY#SV^?#pMr7V`PZt}ot1^m!RSsDq9%D7=$`g|sA%%#06O03PgnC;$Q@3_8O?VD4YQQ zK~@9Hb&2-$mK(8#@V)ehD6j(tA)N7o3C`0>8)}=ht#-z1^yHwU@q2Usa-6}_%VX)r z4SfAl#_?tEFO*j1+~re|NDF(t8WtE1(0g01E~GqJ!`z?Zu>HPJADgvedjuPS$=O! zLDjO1^Qu==Cm#LpVFDR9q^8@LXI4)0*+=(nE4|e3=+^nbT|S=I?}O2NqL;+Idshqg zUK;8{L3?P=T(Koj?AY5=oPI&&)NTcB#6mfQT(t@x(Om%4eBDRpc!b@5#aEuj`#X<)!L~SI% z8jSuJ5ilWhbl9X(=vBJrf;fYow%*)~=QAMlSJe<*SZdhaK@bcsqMJl?5wG`<5T$k& zvIiVc^N?s32C&qe$Ez#^{ zsK#W4h0thsGb|2zI9nKK4cDRyUkploR`z^04vQ!a7x(%k@8r5%pXpTZl~;+05z<2@}m7X;2Epl_!Y zV!J34Of767%`BLJxb6%7Q9)34rV0VB291}W=@!$UOaI>b@Y7x@! zF^P`7cDx=1fq1)MKM}2_B)#^Qu5(c@7cvCkCK2WTUp{|qq{HWK7eJN_^K zJYM(>dr1hHw*)T=*6^oBqLxX~)|{>Q{xTtRVVu9tk?=KCJ6T&cP40A2+%wC?d) zOJi>X(SQn96by`|95SJwE4|lqaD`V4{$*sv)yN{7R%yR1A0hIo98WaqHM?@sb6pfP zFMkBuk=r}vv}R;fltH*1>`_7g?NjbKpzDN;zo#s!;^H1wbQYBC6#R^ZJ7+=8dX1dK%w05Td#;xv6nfU0pWUt+L zV>fa1@2eE>qI@rVoh_axW-qjX!jO=6>YKoS!F?ODp~IX#6N3-@^7KpJ-8t^(i=NlZ zBPxd4=TZcavlvvXQf$5+?QeS&=n}@#Ski7IWV8U{z)d@Mbc5L}ioufGAmmVQin zeCg*H_MbeuGM|0^NrKL=C~B0!1imo3Se;;rRWD_XVED8Hueb4oD81OGjB*Yb1-*PR zHjO}*h5D=C83762u$qCxyYB#(Df}kPnip*#VIU^s5@QL}o-gca?7Rdg+kWL%oc9E8 zb)%`SS=P2&GcoAkHfsfL?Asc!?68Y<3-+aiPmf9$77k>I#d8qE45+nQXwuiX$^kg^ zXUHu4lD6V5y54x6zbXQOy***8rF35d*9Ye70yF{D6RJ zY^?)l=~$gp)C{avPtbp?q9Rq=f3E;4>XnH&mPVx)=1;RakXO8-^d*iP|%=}f3Vf^-r|JpkW!vWDAX|lU;#ATH1$1IX6Ln9 zwx?g4=hOT)iv{I7dkO*L&chp+^(%wSS}J6b_B|ip5?b);zcdx%CTo|IR!8AZ1(ZJS*XS2vkm`-$NB>8Pdl&jL~ZxUj|_&ye}jN?Y) zY$U^KA?M0D?5h^S4!&*v+C^qEYq2Ic|J5gT_0_=}DUxf4P}xU8z*e1g$URC(Bz$a# z$7Jv8f;bWBT79M3m4o#s2bw6oZTs--4Tu|x*t|+_$YqOsugXdFp2I&b2!-5Efp9{h z7<>Vm6x9oN^Oy_c>W_rJ8w`m+Kn`2o&nSFAG+M6 z)+(Vk`6ioj&A5$nfurE0GPU|Bi%%5)lMd&scYHiPWBG=pAW*{GTVgL^Z_E<7K(Y2A zn8+|i5#5Fh|8i}wU1Km7Rz}r^o3VG%u|B63HQHSHBo-iZe6GIweTf^N&uV~guG0+t zUd?W{^zy|p=tn^lDg+fm7K)quLDpU>>R_0ffvn1?-HZuqp2+>8ed+sw5`06!6+nE4 z#JJ;g0suOxUdjuprX<>^VSxUQHIMUkln%%@Zc_?IE(oC$z6%h8cBvq7;bQIRihgHx zrT$#*(lrjm1o1v2JUUX{7StK_1xA<9GWe&9*2hFL%O0?gr^m zIlPF5RM@hokp%kiuznXmZy|_l)3+8()dR;EkxpocLOZpx(l$tYvur4K!Or;Ycb1&|j?h1(R3u+T3Vb;G5g?Ey<_E=)XCvJNkx1f&K z2*87eth1GXmTFEA9Tg|}N{Eo`rhH_YA7=Cp_ru4+h3Sk?k?Nz{4g%v#4~y;smHI8yqz$E6ni8W zw1#CMr0V2^&nJXCM+T&ziA>1WLJQ57ozh^Pd*CU43s6UYu}Qr650^hl4mGhw;0fZh zT-Ht0>keH;{D0Kuf5kYl2MU^05l{Ss)1zwq2vZ#{_su;#rq+9>40P?UHlr!e#Td` zySy)uk8T#8H>eW5y|0;$Vj;qrTx7ujjzg6Och~JY*d_#}ha1#(zY|6&ZmRLT6XHGc z7>ePh@`A2D#*Wie!d0pJ%r-(f$kIQ4l}d1TPE~OrAm5pr*UceIC2uql#DnTX(h#h^ zz3QN7m34@N54ZMSe73QIv9k)Clw<_%&~g$k9njvNzW3XA6sB3Zova|`O}Mq&IA`q0 z$BSvWOQuDh**N7welpD{S)DqleD%3pQMSuUmCEo$M8*iscu=kn@(OqnS1g!3ciQSo zMT0|4E+B>{vk#|X$A{`hBxh8lDJ@gpy(3^gJ-7*F5QhyD7T zyu))QHx4DnS~CDz9ksvht9^Xr{U9iATc`?7>H5w*Vg<4H8M!))aocF+6|rUsVnae< zDYFNW_K#FoeXqODI-J2o@=#)drjYY>1a>2ZQOo2! z>&_hO<50V!@MV53Yz2*&DRdOugEx@2t+!k!MPt9a{iW#kByrr^RV8(SvN zRr&Dp^mbzcFRwjf90ZL0(S-nLSp7U)VDicRxU%q%#dU)qoip;t4WnN;?Ow?eYUISG z6<(+B<;jHku++1d_Bj{_j{ykJC+J9E^O+V(7Wn$>O+#*q2Pbb6C|)>DZ1{QIzzzTU zEV##9=wfz|sSL(bv@<<`|GbInxWUyG?x<__)V;q_u}~l#;r@1`s=(G~cn%uOvSqSF zAgd`T0oK=SnUt`7`8Q&?&c^Lk#XSY1g*(G<45h%?sg~XtbTvd7Wp*peNSC&Vo z)}j{xq5}P?n*R1p;Kep~_3TDCy%mQNWh;KhVb0JJ3qb&R^KxhZdGbMQd?Q1EZ`it$ zFct^d1Q0+N8Y*yiT3f~72SF513BJS znbj@Jfn=oxu^~Sv8@kKU#qua=|D=(cgQlqM5nF25?CEUcYskJJ^K%VXjrQK#&%$gd3>2kO<#DXGFQF&{&km>9A`4zNchEP z&)OP4PsYoq|ArGI76yYkDeO{yl_gv;Le~|hB((ir&p_#Q2r?dP0R3R;3-fAYt_IV- zNjH5>4b8=ws6uo?ra<)7ujlM=?pki>I zVP>G80)n8so~>s7C0Nt=^lKl+!;P~HCwbjt#wm`v<&cxV;5@+3?%IlUT(m%aiV7Ed z-nusu%pXA$b$%<=npu!wDaj92vCH{PiqLd`?SWSS_95Vk=n=rgMYdJT^v~yU?IifS z!G8t3S(Zu4Tu8)P>LT|wh61}0rQ_Zi)2uVZCbG-fMhL8lINs-(f~3Kof@naV?Y7EI zWNASsbC*HcY+n5!9D?i6;}J^xQI$nd-?Ev#NZitqIfuU1ehM0XsuW$X@?l>^gsWrW zxL(SLkct1KPGjmDFJ0Kr4f>-R&xbi4ZXjg5q{hefq&{L7LNrcNgj2(X!8dTdo*v+8o{&eo-GPvWKAuMu8%(7;27G zv8XwZRw2?YtGpU~*9Vm4W*m=LcTqf>=F8m3KaCQw=VkjN* z7&}1wW!LlB%A|fOn}aQC&>`JzYkq&lnTBYO+@hb+u`;- zJkJj#vmH>F=?GzHUje$kcYIaYEN0f%-bPwI*e7@n#~zOF*yrQBKQS?9OZ1DH6@9x< zXu7e?=L>=}Y5^%ZmU(jDX>91QgB9lK-GlRV{T`6hlFy_Z5CWFR`4cwJ*K!@TC21_6}w-07^ zuSa8+Stts*vrQ%}UEzh@kLV0r%*u3eC}fK9Dk8APQzOaX;2AG0DfCZ3{v%@7V5?-R zy%zoV$Hc4vIbuD4gJW#?T(2=s$wB1UHplmkKiD|2Ll-NC;8dvh{YI?hR0_cEb#Vg1 zHr(4w5yG2^L2IhB(5~HkKh3-J{r6j}{rJz*2g4QRlE4wCVxw3U;oHyOzv5(O7!^rQ zZwty-@bTqRa}iQcJjwRUZ-+Sm*!v`3YGph1_S(}D(j$(K8@k{vG6o^2mFV%Xm&ajK z%G#SkS6^kR<_v$#=ZGI`a7r_0k%IzetVXUfhCVb8Q~}kXs^FPPjoK>q`KB44qzUs~ z@`2BojZZATQK4%J^4_B!6YQ#Dv1LkDf zcm57$_~CHIkDnS&B_iB0S`CV%dFR0{oo|OVf=SE$7`|7C6x7GRRg_Rnp3p}T(agSy z`$Gz-eBYmEb9IaQw#jl&Qy7brhNyim0iiG-Qjg4Tdj7t?c9^!7Ue&pS{&j!<6YCY2 zkl_qy^x=OMv@SV#Nn#u$Utos(j`TI+o4o}raik$hNm{Uh@o5ZWG9AP|U9&%Xr1gnG zn1ePL!Ggz_>iimNs1N?Gr%)Q{f*11KD}*U}0+o9Ak~m{NuIo(GH9OGt=bFmvQ-(6l zz>#i%uFGlBzfsGtTU=m>BEPw<3gm&??k_$7IB!)x@6r1`q+;!?0PlrRypE({k;s_JFK1f^}W-11!e&bcMstcr3=VWLB5-B?Pl+h z3*NHFr*pw>hrzj-)$NMB!tXncOw^|;jK)4J?i*%CXQ4rCZ>XB*Eb6b?Kt4JT0%*`S zjb6+Y44}^DHH#lOa^XZd3HC>}JG1ITBE3lKdCnUH=L>xUWkD-fHSb|2&S$5WW9YtP zHXo+@G5Y94>!@UCpY)IyO`U+u=qD5Rl;4$B*4TY*(qkVHQ_K2 z2PhG;K$Basdut`TM7RjPn`N9<-tY07x2d|zBFP;JeH%0<1FQY%PhV+*lrv|pP+1{wXyXap%nOv9w=1h_)5WZD&&S?pFk*$dpxT@>(^F8EO&40*k2ug! zahTNO5DbkggLz*`9ca_w6d8=bFvpM@X_v<0{p9_WJwMO-?y7{jPOemg%bTrCY7ng( z=BL^Uzwa7%LzJv*i=2x3s^FViWhpS;sDPRgtf!g0KK+Cge=mRn5rQ?E9eY{$!{>J} zPhjNZBVl)=MJN~|2p|EciJS9y*1#lVedR0_jYvsynitjXDZMv4+FL%Bkz6a;3ToHG zoLs07Q z)<-og6pK*4rL$K{l`0kZ<2uwJZAKK9jD#Q{VWFDe^_q7r-es3B$A_n;RR_F-MjG1T0 z2N8`Q^Wx)1a4rb_xZZG6&yxZ5_at0!p*)WHKnEG03@&5fnE`eGBCQYc? z5fbrhOl$j#@+`1w8PZb>%I4%%yLqwq(JYEHBW$)Q4%*H&?vFMjxFX=wo+t?7S(XM; zT!+j(X73Fs$QW#2_|xv62ckRj`zq~E;p(0z`t}2Gp&+AgN#8vJhJHrzKveh^A3L45 zoW}>Tzs%Gw=$30g2S}iqU`}~S8D|bLr3nygJm7rf$;u=0glz`Q3`s^h>Bo28RC;b_ znR(tsy*jPtDh9bnILnH=d|4q|7Jwo%r{nSy#WlWEc@yiN?G_I)M=2`?KY;7Ugm7NQ zDA*0d+V>n$IvBOSE>cpe>Ckp?Z#QXN-g*PNBEg>Qk z6~rZ=-NCw9%2wqH+px_1#_f%EMb#@|>6i~f&Sgv$r;Cm*3!4G|V-oV#N2k1tKC%f>1{`~dXU>vMiV!tiAW-BQV+UaP*;S_WPs zYu$jqozCatAuT66^hpfJ|Svt`sU_`w%tv zeiR##9k16u$u^&c;%h*E@B$jbXvt687V{N2p z@9m*8Y(S6&0TBCo0C#E-3+$YkZ*Q9dKiff4AUJ`hfWarfKtds=s8hspTOj;mdgSil zD*}lhbL`3G11en)E9GP_YF#vEEE*u~yF|1nACEues_z#yTCov17?6hGOG3WE9YjAw zV|)HFWY~7USbG2&SUU~#?;ed9tO|t+#A`yuoFfyE65jj-P0n4?j z&aHjl>aOY9<1I@eNb&{NU{)xesY4|nGmt!SKZ$SzG5c3bc~IR;Om@kLW3qSmqE071 zexfQeF__mSxVNw&>8y)V84voJzJ-VbCsk(E8%}d=^zL_B=EQqpLVo5-Vfq4|(=;I$ zeipeSd^MPJB~Jk-K;d)a0F?qR2UcPa@dHP@w{o#`^iE9LxZCf96v!^Y5pNlQe&oKx zw=C;X6+O@ZXSXu9pY6+*ClRn*3Q>~(#L|My=#aX=MZaz7!K0PeDPmepT6F*2+FyD( zEf%pOlx%$Qa)#o_EgVtf9yd_D4?nxVuimNoAqTJAC$_tgjgksiTN(4QO${i`eI?)eFSi`%*URz9k)3)D69E51A z4z2lq&GQ+BhRcNr>DiVik}5~Tz=0Eifr&o8W4_p5jP?!acSx$h>mubLUvu{SO%Vk$ zpKt`uzI`%9M?_Zz3=l!Z_l4@1_V*9@alIwokAucd$;;~_Ns(r7)R1qep#T~a@m3g- zY+h@qH!oOHlTlDX$o*y9_*~5I+kG$(QaQFZKIFU%7(GmT(0e zb0twmUmmyrrRyn@`fNWIy}F!QoyHMh}&N&ImI;_ z++_yoaL|PT`ULU#d5}0DCdT`jf1EUgP&~w~Opypb zXkljhFB;3t9RXoDc5bkCNVv&|H`2G??zSkElVbsJSIAsz6&rF!VX-pkcG02QylrE% z2gT(Cxc>4hRK%MjRK+?wC&bwQNMY3r1=sn$;NSQSe3Wbd>#;H%h#LtYoO)_d-m~*R zoTUk_h$zZWD*J715N%ejy`Ok!pTf7G>mX*PENMBh^w1OepPYaayw|wI-@$r?7aje- zdxo!7@C8Y=t8TSr=|L~!16o!JIz+4EdqY6p3Y$wBBSpIEAK-;Fn6F%(72JIm`9=>U zbRC43zlr-L``PurOk)ZTt(un$Rq#3+9AO*Gov)hwVz4|=H9za>ODX)(?hL!)n{&p> zzrp4%g>=fp%i~@-9v&EZ<}endqC-pQdEo^Z(kp7hSSCfJ&yzTdNr>L;O}By&y8+_w z)c5ZXG?JhCMIW2DqH}i<3gbi*5_O)_z}@fwz(BL=Jmnwlk~vQ@F+=C2b8ew zjZniNSXR}chXX183r83i2dp-HEqvMVcYnV#>%Er^@gCMmEV!*=@WNd#Ym%VF_PB8k z&wD^O?gIwFK7-?j`(vBT8HZ<=qnzn1#{GhhSkLWYqGCMS(*((hW38(1v_1#dCAMHm z!%Rn%-KFSXXxJu{?IRwGf1$;)8S?S}VCa7^r+=X^MH3AExBq?q#}1d?eurI@s&qK5pN6zu^D*iPGpSNuNc?>-vAV=ij?g zM4A2dVP>xVynj_k{+vW3*@E<&xv1F$LlS;SKTER^p#q_Cv*E{4X^61N#{vvTSoazWc>t z_hk_V0`NhOZq|$AIg`V`@He;5aMa9;+%AyzN@ch&n=3Fb>hp0A7Ca2NT?zRYL$ zYJHz*1jN|k1qs1M5tKn$=2R8a7VdCJW=!I`rp&{6tqpYOV9q$wOBb!z zs#tTOp)v43_uqc>nd2-0mS~*8iJYx)Tmd(c<+4A{XOp|)We%-M4y_vehkyL@&qd&7 z5RG3c1Tnw+fw1>z;nmkgykT zgu2YozaXAmrBOHZqkwfP(onX(^{q@XpL_}332B6MN0YE1Sdl3J+B=@)`5swB{znH9 zCu+cC*{G-96_t$m8wVD51qrb1! ze}0{?H&%V@D)L;X87VS=@Y8Z{R2Azpw2QS-+=bnd4P{ z)(~^@mS_--c#k#QPe`ff`qgnv(hPB1N@XtHdaIed2giV`p0?LYGe&1K<`b&YVGq5A z$G`%sq8so+B-~dj;ww~I+H4B_07up8C%$3=f^XB)HQZ{GSxRKf!~kz1gq7LgJ&X4M zM9Ki95MaYox2{v`E$`{!f3sllAy}2cYf|O^h!%|AX{8X<9v`rh>ZqUoaumoqbo1zq zpJw9jY1PpQ(HYTrs?VLR2@w@d&gU!oeHL2}h@rRC@Oxh4VFa+){7l7%&|C6WLSDsN zV9yN%zw;Q`7>tyPg%oEfe!~)B4S5Bg9M((mc+8?-Y7B0gGarLfS(A9qYFWY+ImCv?{5nU^#TLv(@lfk4g6D@dv)7mH#Pl|b zX-;4D(C8A-t(Bt!4F;Ay`7EUAM$whB67dQ$?D|*`{7KZE zlVkQ&Y~--VvHdw*jQ*Uyt4lfL`If_lj3QXXyFuFP3hK8(Z8hInkEX9Ljoocyb|V@z z+gql8r}yh;>ZrG_^1qv##hb`=>DT#82NLe8_bW%#qZ$FwR~(z8x9pL{EF4qRA9w(o z(Xo=@@Gv)5uxp+<0|@9Uxo>WTzKcid+5{Nyf84>>bM24fjQ8KN2(M=vA8SBJ+ym-VpYo5su{#;m^c_@7}4ut6H>08t9 zk?f3Q3Rd=A<1&pNshC_b=Z%y=j_O@x=c=U<|4Nw@JKM5Ah}aETqjsC07rWgJ-BvdU z%oa#MUu+5M9YJeLfPz8VcTEAX(so1O$Dd2!W~{*OVV$$XXyCR?w7;4<+~kMd$Vqwm zwva%#)5GEyNo=3HQ=_~C8sf!++)(2jWJ-Fu7H_(n>vE4H$zJ6t*K;}+tQcc2a16H( z^R5AY-WG=@fF*^F6o|*oC8vgEwNoe$`2}5sfs52N=@eZ#5Mx# z2+z;O8S2cbxb=JT=OA2v3|tbpf9_1HXl>6B8d3be9S+JMkC>i@u27CkITwQ<7&kr*C@%e4U17e4;mWY-isHjV3H9zMljD33Ym@`!D z+2eUQhuZPDDGLPyIQKoNBmFFo*2^n)Bqa}U8=R@SON}Ti;FD~kvm=H09voq21Qop? zp|4ZwR27~COgZZ5v;F$~JYrBTmmV2UKxIHu>^v7-0DGZx8g=)PF%s z_KlKR1@H!gqf7I;0g%X)9&cafbLB!P4HJ`oyF&P~635R_8n-vqr^JV>)2e{>SL9Il z_mtftn_^^X-eutMh`yc(=HSTCj6OI1H0XS`@-rTD8gl9|zqL$jwP}0GhNJQ&!{`hp zX_1BqFs{@Z-oM~B>aLO*wP#x?_>t+eam$(3f7K}1j}u0Yq^S*HTeb}Sx>TnVTc+hT zd^2{tY(WhM!xBTeCjBrGO5z%MWRvZjf+?rv`W;rZ?$lmV8?W>vga#Z-J7@=Vxl-}H z`8c=%tC;#8-Jz_}q9%L~xjrFIW}$HG%(xlBF%WuCwF*?7R;RHdEb2msv-+d;HfI8M zK1HW~1MVd0L)f-S%C=2-nqZoVgJ09lKb|m0u}#X6+X!dD=tgN1;--)`3qn)lF^amawZc3n%I9tY!z0CW>B(5;eZINtDg!fhg^kxp ze38@#Ac{apDwi)H{#-?0c{REC*rjzi*YXBZB{zN8q-Kwr#vSBvdqc z{5d}nMajf26FkEF=`pKBrV;C%f#5gyNOq}1zRoQEdK)G~GDfDG)wC_I&29gk0i?t(tzN79%oo`IfF?>o;=NfuaLbWG$ikM8o zBOh;lBZbEAdGxzF9ESD)uGDYuyLNU3sMJ$C94DW4ARz}c$Ii80$`GOu?!21O{yJ;Z zq-l1Jy0^%8-f?+(fglzG1_>EK0H)Tox~Z2Hd4dE*pSm~CZ=XST2PK>s)99^&#ccKw zvftv@lk>~>lxb75}- ze}Nu&lV&?t48#+@7q)R5F9u8p3!fH7luoC3jvNrBIjH13B6^>$pGrMeP_4Yl*6Ufu z%_sD&m0d#ExgWq4d9=a(IJVbXB}&D_7eBgBQCmf&d1nO~RL|bmd%Z({TxrR90aI%9 z=FYg^(tA^l-klS4R7q(;;W#7CSbBmDjIWsQb)cC;6blwbB!$7Nx9ahp0^G7c`sV~f z0y|t!%6sP*Z7Op^E+NW?mQ-P3m#@eIwBLIA04-~1=b&!?5FZIhPQRE(7c%<+nSv$MIG0Rvo&1Xk7=4)ibEm;?#h&*KcaDsTf*t zax3k66&PWI?%A(XuDY)v#&QoH?5o~J`USD>VmppEQdMGz$cOGW_z~|1o=8ycR{mLf1u^=p(#WGdbdPUvpxYq|U?g*On=2ktC$gqL8_%I-~|M4?|#0KAP z3<^U$DvUfHVFw~J6d8^3Jzun{dYM?`Z~JS9^oq#s@EVzm2Rc^`ScGX~aPqn{DFlB( zYVD8$h@66baz%jM7Gl-N#nUpU_>RzzeI2V!Ag?nU(aWQP8UZeaqrTr~E8vggA+^ch zCwNXlPs}lE#V5cly!e1n!|flJ3N-|F)|!00P;Wk?+w=2u5)5cCrpD3&Xy9@S4cXZj?Pw;ek& zy6xSxk5NeYs)hM$fO+;*zkPBI z*2BeBF~s1z8E&E=GKL52_<>LQ$&n|8nVk242Z4O4_2=;DatZx>!n3Z!;4K&T@K&zo zo9K;6@ybbHCv#sle_~peB#A<{FplTB-zJ9kg6e(UXcv|9{VRgwf@Wt$J&pOj=}M^; z_X}}&cu8iTy7LV@j%QBnUUxU?YrDUX=h4Im@0O$*uh~$3mJA z_omP1-sQ0hg#8eMJBA7Lp(B+z-(-$@>|b)r>y&BnOS_q{bFyC6Q68tjkCKw6wEX9? zo9pfE^t*L%;q$7-ab7frK3a8`-4xM_S{|z7!K3TaySnJVR=jyB+{vNU^z+U|yVk!g zuFB6m_c}J&y}ps{)O;gSX}bk&F(i~Cca~GVN!6pfr5R*2iNN*Il+fo~JCjGXUazJm z&erFbt@*yzp0Yi|#?_Z@4Q{>kVMWTz#zgx)((pv0&|+JQtL^>8AB;Ux@fY-y1y?X= zxY|x|+Y!wA6%{7tV_qK|XT>&Snzjq)mTk)fsw2DAlc)-JZ8b_-<`&b*!N2_>RmjBXVO3Sy_=$E zjzUV$yBnbpWx5i?>Up?6jtJrX_S3aB7Nj>M89((tqzF$VJ#858CsH7AVY3$R|I|GM zE}hE0MUhr^mn|i4w;T95)<6>O0l*bV3cPhUd~9X$HcJV>?k{C+Tc=X(J+Z^ZFZW9X zjfD9Sr2>Y5G#UnL8FkjcgJ}Ur3!ir;U2zAyLb&Gn$g56&dyWN34oxF*!7y*1$Zuyh zu9a#uv?wvCH*BtIzyd#Lsed?p+j_Z%dTn9<&FT?aP0Zlw@ z-saess$-VvStc?M{d+zKH&z^&J347@+t1-ckM1;12gQf`<`SkT^Vx6WO|eg?K|+vkN> z^vB#f`UWS7*mBkyb`mccJWigvmO5&41(upj-QCQDN}q;PDO^t%y^)E#Us(qV$yYU$$-=3*?q5MqzMgAg=a5# zuM5{(MKS-JTc2m&hE)5bgff65DwanC4?=V9cho4i?~d}v{(L|}HB?}>JWFFd z*~ueZ-uJGc!$aM!+Rj_1e(1}t6mkEito>`zx79cQjTD%RSEar&9bmq+6sqCj{LqhC z!NyzRq-|?i2mb-vZ2$f3HxD!0Wp!5GqOKUL@b*>6z~4FVOVtXfckgpW;|D1uOld8; zwZiF2p1g?oIj6?U%-4_@w>Dxv!uNZ{tk)7o;2npv^&1>tdS`K+-4egLROhV`TDc8B zKbuwXEfDYU<3=4O-JTF!nPTeR=!rkQQpjLH;PE*}`x1}D7X&K>vRzra(bu#gu3z47 z59RMUP_S=*-RRu+u7~rJ?2XZN!(oEq#HX6t!%V>8<}&_d%6rSimF(Ib^JK&Y{Oevm ztXj#akGIENzQ;a)QJ)KU!F`r|c+?S*9ca(5-Ot}n+s~9BE%`}zZw7_heeBVwr9Tqo z_lZBBWZ#YnI#0ZOyCJU30fP%i&sXXV611|&t(NPI7lz@Z||OJE?dhf8kX zC})?^5>8toyq;AJ93a(Pdw-@76|VR=K^A&_SD$Gx)(5u=yRJs>GmP=*&S&yE@WLC* zT-!b#GLPyW{9f#IpW8+B)VM)NLl*iOj|ZI4mHOoeyt6Y;d*8Wr-x?b<0N`%jyk!LXMMTT)?yAtL5y%_MU zij5+>7F&=I^s)_yZlD zQ5%YW(FdGA?05sPgW}(N&zHN?`$(S!DP+PSMa*Ckhif?)Vn3vB~k;|ozT(dYYmO-dr>8pc18D4$8A<)z&E*a$jvT#^XsHyz2g%Bb!$)={mY zZ&6c!djD;1Q+t(jb-5{}+!KL+m+rMVUN;dmnb@(SPx!|NMV@RIUTiCHn#1Fx-R5At zMqdWfR@y!6y$``0&ad}pUvgvM;vaim%|KDy&l~wCbwIn*1S9wg*Kx1Y5HcMb-v$L5 za^uYhwx1jzLBw=ETyxF65rdW?&3U$;l&|domrT24jMIZ}g2u=gsYS)ib>lT~=km?p zuSC}oOUR}Igs4w-Xt;&|tz)wL5-0h$hbs2tlkxG~w&XY*hv)B-k1Bik^2P!A1zA0w z6Fh!W2l@3M4-V=D`dwCPEaZ_v&YVuh9>7t5Y<^e*LjaugW#cE$5S&+bWa51?pc8+q zpGg-_GtC(CvoIekHrH2mK^4wW>FiE2UrvE5Yjx$4NyU1J*UQd5kWy?hLj+3A!&H)0 zjWQxBisOrPNyqiZKKdbz;6tQi{oAdp(r~l-F@P>=ul5KMHVRTsaJ_P^{d<~Z7NG+7 z9S{p^_I6=d$I*ydLKRI0QAbUj!sHGv3k8yHirIaU+@HQj3A+iOL<4j*#ODMTKmBoH zPIo_TZ(N=qp%~j8w>y)6U@esC^tsc%-5dp@(w7nGJP>zNc@*XohsD8%V1{JKz? zCh`T=ISKkFFY~EZ_Q}HSWmEESI=kOAp6|8wvTvwfK-RO2%Q}J|0fOi3+IMxf1@PL) z=7)>-*rP%2MJf4=SYg%tH_TDNOgqoJBVvaJrv-zHQ-41zfLe7*=VtJVW*PrsJK3JS zxo6LB(!(nEn$3Cy>_H-jP;|YT&hOq%39?H!$B_e+SB;benJUtQwmmcbV~V~Zc?Z#F z(oMTQ?FLPY|N6r&S<2dcgA@SP_g3K zKr^%NMs1`x*@C{S<_#z$8MpFxdtB{d&oZlOdF_+NJMb+(F@6$=+E#SK^aKuI{Z=>f ziqadLrGLqm*w6aGvj`UqxMU}g9>?10y+PBYU1$G}J|-1jvaB)p!G97kuq7I?&V`h7 z&r3pPFh()e({LPbwv)jJW4h0ofMoNPB@9W!fj95J2hwY;8z$; zANCt8?h|T(nkznv)ZKR&M*_XJAW@*rgc%fs6`&LgTfJpN%&A}QuvZRF&Rr#mr+osS z3I#=JdRg8^u{c!T!@tyx?&s)q_88XhQ?lU7X#@V|di@2T=1rl1x~|WklB?`>f5bAI zCs_6HUZQZk$G>-v`anV&aI(%7+O@n={SL6HV)Eq#Vb^6# z7)L7CmXv}geg||?OF{F(uVds?-X({5g>f0NIG7FD9c!7J_Zd5b;Zg2Gb$wV^$8TD> zbm9hP0mhs;K}`V7x*Y{I^7nYaZj&59Ot2kL>{6QP4S?ep%;0YQk*6At_~ z{E6}`HVwjTTRQ`KpASJE{f#iJRah(G=YT^s2;uy?9=z;LsBjLyDs!PIM@ zb1%XGA@7 zs^;aMR*>mTc&I@%Hvu4$^e=BL5L)Z=Lt$$JXnkzVn)qDN8wP)(Nn)LXvOuWOFwf7! zc&_MU?L+J${H#ZjWV!Qrnmapk_G)ftycpUW8pOiyc79()-+pA5pSkjo!G^Y8{tNcB zXLL*&SDHU!Dn1sU6Mkh=TEeq%ljjPk3?;eUu(5d}p-KXbZvF%`yYz0DS$iMDer@&U zDI3LD>}!NI>>Iwqmp_v3BxD2xVa>&>?c{zL1n3Dv}GCfbQT4sGN{Mx0forp8v)3BCSfYi5|DC= zZDU5+{}f5*dk`|3c_oo$;1II;ma`I-p?mL)?d1_G%xOjynQe{^C(6As(~mbqoc9QC zAVP(cX%@r1~9lfwM@Y0XrcSliM- zOZIG{gzK}SuZ(9I`+_emzDD##7b$}j0c6CQ7sn!Q1NaxLKyWjB>Bhg%D(k7Q1oN#F zEY7G5N>r6$RsX*IiGZ;C&-=qbd&>J!J=5U#n*l!#7$ZVq&S($Uwrm$sZUI)aGV7|o zt_i%+M)J>NnP0ET)p*3OrjDM{xL2|T$713))d5W9p6bUnji7-5T)8oM3VVd99AY8f zvZRxz+_NDwkk(2Y*+m{?a|ox?&hL-fwPUB6dV}J0ukb7HEVA#n%qf0-qI2r(KII|U zebmW#_E;rOE?D2iuuQbRuMT!wyACazw`9NHFVtYLQG_4BK3y~IGG=zmT5bv{Z_Kw{ zYIf)5pQWT&(4|p%3k>2YV9H5RyaYFAPi~U__$<5$b>62s5Rd{{-23rL^f3HXZdc&z zK4{m4ubSNR3jE~bqmEn$q=owrh9a`$!lNxjZZYZbJ)O@BY=a*)f^#W>@`E6t7ptr8 zi1q;SO_)z`Ju=+@`JVmWhXVAF|C~EXuKODDecTi@bkmzM(-(by%p34O*dc~!Ap!Q`9N)rF%<4(8iI_5kycWk^H zj2(!yiLdA?K7*@CrX$Gv3q;!#SrOj-2BBMTi`80+nnr(nM>A`C_-&G0&`Cgq2A?Jf8?$l%p5X@)IXKx~`8}4Z`B!7IzW<|Hlutbz5d`DY%gMC_@1G;~q^o%?v29b6JVxi8RI;t#DMCjv!oluN%NA zC;xt(oc$XKt(Dn#=o2MoPTM7XI!>z z=hlB5@wNXz&xf)9R=Lr}X*uq|Z)OJ{KTFs;{M#my*th+Sp8j8tOFx|te_-@@E{MOO z=`8sV_VCvU{W~=JiTDSJ>$UIqAJFLRKPU#32KJ2P0y^n`?eV`4F#U|-zn}Ae22B4S z0@44u!+*q?7vl%;`9HLRe?r?o%d3yu;vdOw7x#~EGEPJOU&bDt_ScSr=d+y+K>Pz4 z=U=ph>)-GH{@ZZ)w-NGBwEMqM@Sc29kjtg{zlPJldnG$2M`k}*R`x)*|An6)zSB-2 z_z%2?{yWBhzYHeB|Ci&9!8rQ+p8P%Dn12wbe@BI2@$l?-?B9(&n)(NehwHx?dxt+y z;=cKRqXGB#>HmK*;No-||9j(&*+&CroqsoS{$Gr#|7uYF{T%(TpQ6aWCzZd4%YTO{ zLy{N?sN($B1LU8B=6|Enfjo9U%l~Y2{Btt-|0z%A!r2@TTFG9(#m|dJz;TmI&w9A; z@5`6Mm-q43lmhe+-jVnBWmW{yVwBz;UM z7D%#sidK#V$3(>^umfk#oWj3g;Pi48aB;4JSt$^#bww~}gn5Q%nD&r(@NkwxCPNVs z6qf0>axf;;urs1od!9ANBy|4Vz|8aj53xzxc8%!2j(!0br_ykw5QLpjgS(~T{LgZ6 zl>O*uFhJl8y}fQ7*d*m2u7n9CRTLUv4a)CioQw2_i}O!<$)CT$&v|JV7=%E#@z0*H z2HCkJfA}RY_rkpZ7XSxGFkusAY8S=1z-L9WKv`Wb>(0Ts&ior3oD6TAu8Tq{SIN;K zCTI&1veNIU5a5NZJF{dzOOuo5oxuV&$A91i;bfq3$(cJZXg_5z69^dD2!B2wCRp6d z1>a@ofZ&e_9>{vJHPC{1a9mvfvZj`ebCDOYKxQ6F24~BCADsOwp}?TwitOwp5l6%p z?g5+~0sTreI#=e9`sw9Jrn-}VE|SNgyquw|Bm3#Rrv@1uLIz~7dVCb0!fR0Fah}PA z)IgiiQ*+urQha%d^zmSg-JolCv&-ntp~p0%V<;M8&4wXC(MFUyY6>J+vs9gR7*_D_ zhQCYR%rUa!g`c1b(9aB+B0D0R5NMI7qf>Ya?33iU3#`zT(96y#36Z_;+zTjB{kn&E zl{zESAznZs=x@CtC5`;5tuNS|6QO5Nmv%Bm^YDe=cwZOKEyZ^stl z*_PpS#N0~r;xr;WSLLGZg8%he^aIawq(6)E9yh<++KOsPjIyIOO+x4G!@>)H% zSEU?Z2I-0Fc?Bha=QhK?z8El(S7hvrP+ZXSh9hKc-J?W)#q~pAH9x$lp#^tMC<;K= zd7o~y9r4a*{`p7=)GxUJSOXITMOCyT%r0%HdPWc^64)7?dh#0>H^v#A+oBCTA*&az zUw?I&3^>qs-Hzf#RnCH69SBGxjj*JLo}#mnp%zQOz6jIj2vPD<2{Tw(-W&najbmR8 z@w-Pt>hNE8y@KmPpid4%=f|ciVK=p_U4qgYH9$2KUd1o5Q7Smr8a7;-LbGBFE<5>JPsxH zQ>0I5qK0OuwH#sSud0!}L;0ILy8x2**aM6RRrqKCxF*Fr^wEJQ6Qqo#tdn(-Nyy5giZVTx~v|_FLPHRR@SM_W3~btpcOztDNpC#uXUHh{*Wtcu-0s zEWdGBa*UMx68MyQGg>xY2M-MRvOIDJ$6WV)r{t7#BmqJ`uZiz@kFsA)3 z$sewQU|gnaYYggi>` z^3%6mvVkRAb^mNHBY0E^R2uY)0CbNGuk;FBM}EH`EPFJHP2eu`3VSP|4%JFz%23xf$^s9)Q5W{W$kz)c}EP3gROP8u&GMdldM&q z?iSyamJMlgyJG4qgIPC`Kjf%`+@)3LFe$9pD>}U4WFMii&&Q)r0zN15%GsQw=>fS* z5Xml)eKA&GJen7z7lUt72|&xdQcX#gv~!2FmpXrcRGadB)eIOC0-5GX0?gr>-{qX$Z~5cNx(3;;ChDOI(c0k@N&?g?cj+>)JWl$nxJyA~hVw4_Ze9f+DxRqO_J_lF zU15w3Pq7lTM8#^Zn*dnY=AC_d@M}c{@d)z>JactUc-uG}`xKxB*T(lx`2b5}y6X5y zI53&A?x(Iubg0pH19du_Q&h&EC6@Sw8oA(g{X!L2EI?;#r z073B0&u^+pd|g?H>FfqTXaz#bs)>?XagZCjIFK>|se{iyt{>&-tafm6m8aq0s0o}y=`g7PVr_q(LzMjR85~0&o2_% zV~Zgw=LDV}c4*~)oI|z#BKIS61~WsnT?-Sq#G5&B-ClvoqF}!f3^^O6#Y`R)TifKu zh1x9X)cMH~i(=g6t0WxKyC&2CHvXxm-X)qQ+gBqM&-?RK@Rzl4YLoM~)>N7I`j`O0 z`6eZR+g8PZH&<39e^V3#{sevbTYxh9sGf;>rgUe@%6MG2->-dgR8+PACpxt=c`%Fb znthtkuEKSAzGsTCK$URY_>~xd{NR_4L7Dfy&%1)}q1H^mMN2ZXr})s6(1tD9&o~hs=U*Rt#Q%yL*t>8i8%iz1 zPd$j-Z{61)K`lRb!J&?h_{PiuY0PRj@Ww0fNa1x-J;@*6xubtUXST#m93*T*mb!5} z>&t2XI320g7Yp0MRKow8>FFE<-PjYvV|){N8F{g4R-gm>=#aR&FXBC~+q z+YWdde}3@-Bl1|y#fMik^YVf-!;1z~`e>RiCM`Qq8xOSy6lbL*gf_coz(}h@1KpIc zNh%N&^~(X;uruU>;4V>q_pwD$G8tPcdpRBN!mX3}TD%CWeLrHkO&}3u8Om$%n!7k< z-1ApL$(8i3T8!3WwO9J3=f(1mWg46zQBuwe+sJjp#RXy;-q3}G=O(HT4em)B zi+{RMPf>KU5Hw)xcbt#w@%PDF?>K3C$B?cHXItQ5fkj>ff#f1@_o0MU%q5#Mu?Hkk z%SBM^RRCt{2XO_0ku~Vn)0S4rrFcDP{6ZG_mtp4~xs?hg)Ha-)9JE zf-6Q$n8GG%fdV$8z@I%t6m6r;RiS3>t|Ljf6$w>F0XS^+!S&4_Xeq zYAZbsJHF!6+l-&vQ=sbT^Z9-!^Pm=BTz5Ac5<_y~ncCcv#Pkr3T~ri%-)a+JHiB>) zwlDyP;c6%qD>jGbQV4gmZ1WXBXD1r3)UPjZdl0VOU6t{;vNTESv@bpV;mRc8V{IRy z<#ty=-Kw{0+uTq?IM7c6p2?nq!+Po_oEJ~Be>jL|r7B6$J=uwNt!v=;4wH7s?%J!q zmz(~5@P+feNsgx8C5x&aM)2(vwZW>=9VdB__lQ8Lmit$&0D)@>!U(572|Ym($6bex zSX+0tU&KuZqWA3OJ=%H;OccMcTh-nKlq~M~bNcOTT7Y=&_Fg1a_8eZZx#Mf@&B4Fc zKa0V~Fdk|7FF+yi`{GJ>SiIjs<8J}@dJyeWu|PpcDrff^X)?t34}Fq>Ti004%pyQ2 zAna0ZoK7cL*A)Z!mUi}mb?Wd8l&O}1-C^2nXUC|2&_m(vgdWr-zgX**s-)3QTez2FsU3K0zSiko`;x`UWp&` zB;CU}`#zLePDjA$tP>c2k_I=xp;P@5v&>#!KiQmurC&8G;_ zPqa>dzPVxgCqSH|w}pl7l#Bl8V@?Bgnv+lQaXg)yMRrH%V#EH&_#wCc*s~~+7yd(r z2vFGbz$cxi z7>Kab)7I3d-wYKzmtwCi@Ajgio1Dg&{CR!w9YlV*t4IFZ=-PV+yeL!Fqo!TpZSfKc zfU(|Am;XT;JBRvo#lzh2Lh|a?rfyF^g8!XLA7>4t>RsOYkX%&1;^8xG3hXgGp0{zfK1#YZ_a2kbG%Rw9iaC}y;`okif;tkihR(~awiw<=^|B6aw>gJRBuSKrsa4a$3H|5BKR; z0<8Oe6kTbD1qh%&&DNli2bsnIbJvlYE9Kd?48*i=ye}WU@0~n9FR)uNJJq0!5RzFq zICl$ldtcZe-3(}vy_VruP&-~!Zb~10=r)1NYSgI zw}X23?G|3W?j6xAc{Bzq^SIwM_iI9BviwnOc-6b(EIeMz_WAGx;u>liw=l6T(3YL=9Nq5g2|gMcjx~x=P@nRk_M31u^?1;u zix&%+8?E9;yQ{Sk*5$%nDE671w8?euH+RMa0LwfOk)L~LCYx^Y-W}#3PvFYrwYyIb zl=8`sd~H7^S78E^+?+L>Lwn+PG`VvDJ4C`%&m-*;_7ccnto|cOE<^}tK2#ohTyUX0 zb=P;Y#Xn@_C2>`>1@LwH7eh<+xXt9veYvRWW`1Gsg9Z|r%j@0`USe5=MFI*1Ki#sh zpoV=t9Wet1Qf0yS@(Mk$H&Y%EG)UsC{7N+~zGkHgzBo9~4bHTIlwaD2KRFCAq&HmK z_c4Dh8C+grbX^$rV(MSunqhoDJZ!YeO-Z>wqTNIRrmV5_)go1xQ|@JK(B@2@%~RYx zG;0M4>w`%yn>;ly>}z}lD1-U+ZLe>sXIn-iDA>f~ncA^?_)#&u;Q2SdqT;en_ZQTO z@HUK0B{r>#hs)<8>^+jer;hn1@3t9V_&0^GlP!1K<>+>I0*3+BhX4XrN^ka|Ot(V0 zPb4}3e|<*#yg-O|+e6tH<3if!JwvEGXebRkES7iWMTuZ@qC=LLH#yWM;(ZR>=Mq?A zM{@G*5xBcIebtofT6g>hqlg_+jxBsB(c$YBG)zGH7}iE{fy_rbe+O~T?k5ybIP*Fz zfb@osL5PhYOwTFrFX87!Doog*^#WMnM&5%N$aP7u7qH3?rttb6CZAe1Bta-7T@+PL zQL_21thPUc@fEmVDeDN3R|D9fUL$8X5y?f#y05;4-uy+`#ibPa%b)-ilh?j~zk}>Y zO(@qCVs8lKBM=PoT?FCjF@80KRjZHWZqMVm8+t8<5|Xwr?bl!&(q~QH^}tLAU5Z^g zJ0k>#7ccnUsG6S#);-$4_LU)GD`ijMP|I`8{WLMjGwH;qSn#v z<`3UJA8fBI?C1{N>k=hQf)yLmZXvq-@r(cyZr9tIkZ1a6eFIg6yTN6()Ad{WdFMb< z93My@3rPH9$>=z>5k6@Qv*pwAR{}~a46UGaL>IB3e2lo_yjZS~A%~bNv>%ur%;jAx zk*ACA@kpszUdn08n2L|NOjfcNMba9HI3*+Cp!lWXt!b!K3WJmr>T57?< zL4{+Bje%^^kNr6?#^Lc^I}h5|?k%#Sd#A%4aqoLPHsB4W`dIvh{v`aS0(4)LK^b@@YL3nxWKK<{@p)fy?%FHP-#zLuU4gQ$klEk*ZeX$ z-XLMV7xD*5Z;9`K2>B(g3&H6DHZ~M>a&5-_z8Ls22u(CUw*Hg|mpu?tM(aer#CC&0 zOYUER2(-WaWAnTx{U?%J`_yY8WwTb2Q@r9!^>hM9Xx+AAX*l!h-{**0f)|r;fX@Uy z{v9B@kZSJhRh)devEbngX+o3b5xS|Jc0Hl>z&=jrT9aNg3{cwd;g6;}qP-F)!cG_04h5#Wyr8c+H&LFm8yRG;XH$Ih=l$E!@d45cS~MV`h_hXbq_4dmyhjLXY0M|8^wOV5Ye%9-jtekp-#}2?yxu7X!r>a2t(b4Py1ts9 znfcSfDZj4{s+ggXGy&~550R-&Epqs|e#KLqo^3=@Y1I(cRrAV%myFZzOI^6H@c4YU z1 zbiAi6irXz6Jh{C+?#Bb>XlxuM9^ru5o%RKC3y?wb#wAv(inr%&(6c9c_eWs~!{Zxq zz~u&o=NC-eE?@3={4v6GPvyiOVDI2s$)4EdMA$R|SG;Q(W?(NogFHr?;6$CaAWGsU zy=21g@2j)WYA@5%#6W!=1wDuu27BD$R+q@ao5-1=Tb%dEm$mTjHr^3_xqEHXhR!-_ zHeZ&lKB_W$()44Z&pr=pEaJdQnE+F|nB*(=0Am5(BncME^C?UKKShDnPUTlGy?Dp{ z_H9pzfo%$Y!->xbdkuDA%{(BPtCZhwYT?}MN*gq+1Jdn`5Szcc{t`!V)liN-1V-AG zTmeq zu3#N>qo1eI|3Jdj()U8n8OC6BB-HSTdP3Hf;6 zo#miQZ5D*Yae*G+NPgPgXwAYIWhkbTfq5S#lBCzBbmuqCt zCrZ;eH@NfW@Fe{M6`aN?zK63lV_GSS9>5Q`QifcqM+M@fu=Itw*i_ENKmmzRjnXv8 zuHe`cx8ZL_syOyxL7cdxlis)mN&ic34Y4~Bl{K%{RxOSgP~ok!DJGH18R&%p<@2I{X`pc|muv_$hV8tM2agMN z!V;|sEvs85|CAwa1v}`Z3h~AsyU$~=+CQ%#+vY#co30@`jHtuG>TX-z@GOhmn*}h3 zDhpJ3&hNPD(+J7p$sn`mA$guwq*Kyq?qJ{{8vaNLv#x9>GkZJ=1wMME+dc|jfiUSQ zakx!99+H&$N)>U9I;9`1NyK=ynx=!Lb~KPa3-Bgo(4Hm+rT`_UsG26hUW}9FQJ(t} z#cwe|v%&VsrFu&)&y?*ThO6EHf2zU*ZGZ0f9*j;1FsJrXC&R(Srlcm z?Vgi0LNdC;%%8op@~^%F&Gccb`LCfq{S=rZybk7-NFph-@^IlFZoe(ff&HKl#4Gaa zy}-?GnM!H7zTid5N1wYRrWAdOQ%3HQmG5N}p1y$u#5<$pUqPU7pxZzW09rpt;Yx(Q>jBS!4|wkfq-0(dRE-CVeaJuS?Qw>b}oEpo@_k z$1~i7B*AM-_$8}4>Tbun1N4_Aec^u}e6LF^&9dMLGs!H296dxqQAY>*nL?EUHB#V|$R3^QayFW7=}4)f&p zU#Q9**rQLl4#{T)thp|s8}gOSM;cu%~Iv8-10aPBm1bT~TC-#&1G{!RAVOxhfe z-m6!ISa0X6e)F%7#9&zrBYt~(AcCUd4LK?+whJB~lmgqHP8d9m-h1$#Mq!Vgdxgy% zEyKAVAlUem;Jwt3JOS&dkt30yOb7h!-;k?GPsBeL1e|4))6@~Ks}znzgwFhKNWN<8vC+NIekAenOjy44}^@6 zCF8Sgu{kj62D702T@I%~Qy;~uUX>Ny0Uy!xT%fSJz-;)kfev-t)57n?>2p+<=mH21 z9f$`Nw86QGG%;!mz@o_f3`*nUf`!Go61P`kxA&)8FMqEDmRMt+cib-ws({iuz%Cpn zO>2DbOQow(mZfPDOy^?SbI0>3_r>};9&hCd$@ut;_XYYEJ>%X-uBE@PA;b)+@POe} z$Bnt7MCdu%PQ#cmP1BwQ6m^oqDFtQL&`hC%aZQ=~WyDN*>V|`LeSJH}K)larPWvI1 z0-(n=fA9nn9e5}+!nUkO@_wr#_P}IU?rMJFFpXb;Yxv37H0*n?4ZT8T4r1nMgl^^pwqU3m8hnM(-0>kb{ZEYM$y)5YT9Wn=1wBX#3+(*Efo~?&9#8 z5bborlWwl)@^LYguZ$E1@!hXt3iKRADJ956B-KA-K3V}XAvEAU1iURA-3t0L`yYIr zn|L!}Gsj2l`67khA9y)9-Y;kJ8j6Q6T4AT%DN1@kp9&?FFZE*`>($3lL;iJh`Zg1@ z!=9BfAm;jfG!3j2C=bQPOXqtUiSXZ13(ZOEP}v{cyv-cvUH)W;ejNcp?2f>jn(?DN zpl7s-mD~v?dwifer!HT}BOV_mM(rAlb@!u5C7PWBxriTrbFy{dh5X|~L4bA;@we^- zv!~yFQQ{drZ>@o|3UHjT>kYT>efoZa{XF+w0)GsTg=gwl6Rts9$biW~8Kr{VSS>0# ztLX6Vw$XkRo!)@qF+A$$F1GzQMUeer4Zdfz3nAuFdT40kV5PU_A^!WgLk-v$gP8xi zNx{F+n>WSVSXEMUA-{{;$9@z0eevME0LlSPb;jefpN;S$EV#f%2l=LBH(0i-{1S7D zKrlb!s&NVvKLI?sBAL|TAv`HIvwxQdG^463#4Z+~f5Eb;Vy2cpei*Xw!r-7vf}y z0Yp>YKd}a+?Pc^oiC6e ztg#_;azDHc$Ti$9Gsf`eBP1gWF}fY;!)Y?o^S*glkP{0razM4q9dRkR_xD#t-0$e? zRjS=5LN#g&{>&7bw3)dA@WGhgp3iT$*5e^ljbF)MXPUYN!Ifw8ll`gI)AJ-pf@t2E z%m_ch4Z|Q4GAIb>ulR-uv!m~DfrBa~o#(+!id5!1|8ZfItNIP(*)A7rE3rY*oB&lU z>pk~irGq=~HObJo<9<{3n??e}zwOyVQB^9sMD3MjGgHe3qnZs7qJAwGoKGUAX1Gw? zuQGuwyOsg>s+R5;{oVsQyG|k!*i1zh?+6Gf%3dI6JUh9h(dP|lk7anpkg1jbGZhVH zsD$hDm^N>yS3=j&Ri8rzQ%#?g6^kv&-1N?W(1(|(9Fc{chH~0>{;EKGaISX6MIuu z%^UGW+Z5k^+^2U0?7OJZ^E?{#K$>#lyZ05!_*rP*0TZ~rL0$c{KF7TF^irJX2PwCB zvM*B$S)**a+ep3TqR)aod?*v(n40bG`GMSI=9ML{3@ReAS~vc6-_tAn1ZZ_|k=;Lg z`jg+VxxUxo&FLvm+RDD%aqG5}FFZ1R{%8i}k=rZ>T?&uab22ZI5qZflU%X&+{QN_s z6Dpk!Y6~gepSCAE3ySzK#HijgNF;Kfce7`1K zR;LR1Z^$7l<2GeAQa)D zA~=sG?40L0gV0oWt2)!39~8<({4Qe~;xpET0B>xMZ)gpa6vcU^c2)F9U>@_H2l`Ss z1w)noipxlhMTg`7-u81#uvqhwECjmDN_09v;q(3>73zVA=b5bL;`uJX*zg^pz3;tn z?;xZA7Jh3%0{OY!$blQN4;qOTpE0UA*aaQJ0<;5@N)?tLk&?)(oVIBt0kf>^#7gSu z&s!%;GKL$EL6qTER`cxuV>XJAC+0*Nm(Tmg6aZpdPQQE8ph`R1#Ecf~^XCgL8Xbqh z6)MpTzL&nA5DVp3@`DNn2Whp?fcCzl!dUgnLSgblKX=bmEXKLI&?;&L=oZgBX3?Iyfp2TTQ?AGD~HmO_43PhRwizV@DaLV>IOI3MNu z&2Eo>x#&$As4Fhnp8Fc1412k5)t%Z|2Ik@BcYp&r5`fMr`d0}=I1~N!Cp)k`<1-YD zLUZZ%S7E9OSv2bxb$Fb`vM}~Qh1X?hf~x>={NRCn`{wYttF-)QxoMFpM_-isJW##t z`0Pd{fS^zdDGSWK277%C#R$8W+vg#aoq%0Gl@BWXa~)474&U-{b3(|w$HpyIs{Lim zG9H7ykE$dlxrYT}<+vc5{@aWvq|45*2x8|rUNr1igUd`<(YeH3f$QNFxtXuPD2m;i zoLVj#QJqmpIwjOiD=a#JXaVr34{xUz$U38z$P>lMntNOB#O=_}XiM3x zLNM+}dMcbzx>9P{TmGEXKm}5W!)Out3J-I~WyBwY^=s8=(i8*a zQ@$K8HYRef0K5QgBm3w4orQI|(9IA8!}gJQJ_qV<4 zXX@EZnQDu%dnfnk)lUYdJVaCnfwZ7lXbwdXp zry=dlnyD?BkDH(Bh%P?mXLn?ETRUvE3qsSGqCR{Gho@m^F6A9Q@KKs(%XNfTev*u>)P2DMJ`QHJaNPdTN@qTBOkqBr_rn|rbE zM2$gQlsmS)p)QsK&M7&UPWHzCNwE*ZiJt~T4fnWr3>f>KO!an+hTGM&McsTI_gW66 zx=#iYiHGcwKWwP%CmU^*`#cJeFvW z+xxPaw8Y!MhiuTGp|9&VK04oTHKr2q{}h3CkZW`u4aiF**-y-HN~l#|0qs?;o6)?^ zr^`7BGqVp^2Z)1Hm>%c8nhHs85a`(rN>(r~)Cj|L>Q!hR7J{lw;WF1YUyu*2@U8PQ zz(!VnkIsEd(G+e=SzqW2Tl9{wnRWh zne=0rnJF!B14mcaoy!8No8^8w+*fH9_+EpXxAz5V=(;JuxD=yra-yk>0Zc4)&>7#K zJ$E}2-=)}`@2}m)y@U6PW-6qq{v`TPpsuPw`0`0~rk!6M?{^ad-lWQ5Pv`&+P~tv5 zQ0qM?&1k$vKxPu6iF|us4lr4fgI0Zi3jqC~1sd4j^Ko%D5f=hlFXPaRE7yVADBqX$ zcL>NAAaL1KZM@KAr^uRQ5-Nzn!-*f4C>7U>TI!1v@9GG)fT%j4tvASk{FJ*u7Zf9( z@ks3f-*lb2K9lISpif3>p0Zb48ye!MUNT6uI@9BG6Op-S!3MeFRkCp%BzSfEp%NB? zyZ#nNpMKARBRJtq*z-Bb;6H~(vx&`z)hkah(F;oYVL46Zaddo&KJLI^!L`IDsb}{G0lTiz!6!vg zGmi3K<9jp@?RSqvC#dwpY@zpRwNz|C(VIf&u?v!?IVqRA81%Y@e}NMkYR9L@j#IFg zhJdlE0eqTNkl;FP-3o==Q~S9eGt=c>27l>pScdX1!e_RIUO_0-bQ2!WkJ zZRv04@13|j=lLxfFDs7W<;G_fF`kBL+)dgB04bb!{`tmn zzYXGidN$}L3^58+bfTcs`%#`!A$y1E zjYZLSp=F#6$m|DKooz~2mEJhH;haZjVzq{w`2lwt4&%MrVTDT2oa>AGq4Z(g>zpTe zyza%5M3$FSMV(NKaC_eTao^7K<6?^*dH29m5qS4d_%eSac3WP2&G!X^BWHV?W~X7j z@pk?}&)eVwgFAzWTVQbI6SvWjLm0SeY}ei1`!fCQKyG?Y%z*lR8{ZrSx-!aRCO30# z2%g8V?7jWkh9ZFxT`1}msF&c-(YSu%U91f_pTPlG@%n_$DKZ}kgK1L?F! zx@Mr}r4-r@`#~4=0fgm?Z4!t?^AoAo~^kb9o77?`(z$ zM+~p2a=znNF~wSbz)Y3}W7~Iud`*oZE$I1M!iTXRrrVHgQ#_v}z~N&$a3$e6TP(#^ z11Nn25)*D7jskVfDK5*=Y&-%J{iC(uMTNPRA&ioybN`2Zf%$vb*o?D+=YM^M#)Yva zv`RWQfUx1ic-T4hEtw(|O^24}_L0-lIeD>q^*mu-{`}61>3pMgzjDnP?tjSb@U6ZI z!aJ|OwekV@`m0ig`?!N7kksCaGkgV{7l~m`wGo4Ot9b7xQ^e==7Ha1uYA|LJv>!I2 z>;5fZ^wE$U;KC>)mJ)4DvOB0m49`y-?nEkTcYo}Q*TOiI)soS5u({!JGq<;TgDu2< zt0%M;z0}+rK&Ek&$#mFz9PU>xF?sfO<8TfIAiZ9$MSEupbw{6oy=fc6z?S zQRMr`G7tvABe`6^^BJIiZ?QZ^aNq^RvNbpZ{C@i$v)^}J+3DQz_!G4R3lRfn9J5nb zwJ}}HofU&*ucy8HFh9Q+ShgFE^R0o2{pQi%#J?@R#T1Hs;vM`N0>MKe@E>S>*yzE4 zNdX0lk9T^#?cD^A!!E()8iRRt>74+5_yi3C3m<1@-vECV>GKuT=mhgn|I87-!7%v~r;eT5{pn$axsO^-bkXNHhXR|< zJ%zO)tv+XY8ypu@F{8^^?Ie%{e=JT3#X}fw=q*bH)Ija#a~|;i_Y5Z`k1*&g;tHU~ zmS1^J_e=3aC{K9sN0sTs?O&bc;!85>4rsYICz}A@z`MSurw!j`0cduN8iu2RyTQHj zEOG*oO4|LV1EGRR0otxHjaSddWOgfe@ZTI%CiVsY*sr}Nl~pq?0hsF~JQncEbAE+T zQupB)G~1jLvc62)(`z|I|fme!NgJc&bO`QbyGKgm+!;g7Bc6T7`9|>sXQOfus;|J zs%t>9ivg_Bg1pZ5-+ZHT(GN@g6sqH2>In(>QxHyGFA~tgd`&}H9+Js6_r>qSs4bvL zvtfnyo#tx+Z=ROx?|9x?J-hF#)Rmeec9+O+Hu%u6Dt>Ff$J6?T4rDwHKbNIBsj=u$&YDZI&7Rd&0kWM zspJdJ*%3d(glJ8iCeTH(0NW#juKjgi#SgjL%$5x?Y(>>_IRq$d@UnsgIK)Z{8q<3f zomO!=7L#;&O9zmnt2F7q&m}sab@3+vXDI4>iozO254+eWAV&vT1KUQcB+4ao+B*PX zdgoeUp|e1*7C>tPRDc$M1;=* z%(9YyXVOV98Fwc3r@*i)u3CO}GZV07?h615e4mYKxOF$DPS9C+%Kn|4WcF7%yI*N4 zd8Zh9z`5friJ;{CSm^5m_HbAbb6`K}t$x_r@6sv#2cc(j@mxWo|!5Ign_ z{scDK`F=9Q>@jgL_71?r1IIC14pD;KXJw(=ki76kf_BY;W5HChe?j8m4Jn$3&vd{E z@ZaAV;{~{7n!Ke3&+!^IVKUu|T{Mt;gH3Rf48Y1irS703--p&|>F+slfeDw?%AC-M z&Sh}=N2rg$h|BU}qOJ7~WNYBvN{&Q5MR%pZzI;5#9Cs?oHW?-rxMfi@c}4J+2;E&a z5!(YZV8ksvn)r#?jGCXi1y8+{EcN$TPKTD>i}y)`8KWj`ivz_{fRQI=e^{XvTHdj_ zVJ<0(IZaj3!w}S57!1h1zzFllcSH3}D|SFP7{}~#jL4erL2LFgM@%gP4;N@h5 z0u1v+0wsepVzkV@*d5tpfA*(q6chu{B>^Vis^{*z>~Bv@Rkaj~As|mz(10+a=gQ@E zop0&{1PRnPozt|}&hZbJfuaI)EEFsrjx+OVE`bCw*a38CX)kL&drT~ry_apK=SKxP zqC>)%?fq&ceMBLg1QOIeCn$UKbXqsW0RP=wlw;@kzH~uECy&p)bqrC+#1B=!h&O`+PO%v~Ca{!DcZ#vLSsVC@+rde2GF^`WwMUREguEX;po-BRO z@1SV|I(vnlfLw&Pbvcel!OJ zYu`&Y*zNJGuy(7?m75qw2*^_bhI+%z@nQ>(fi z>qQ-vHpi7D#W`z-SIM({3OMG~-B$8peEaV{J%PAV&jcq$YQ0F!j)F1p=EA9+OYPSK zhD|Vle`KChiEDQY^x_4)VJ>w;L_n3^Ykd(}E2ORf{OGmYVidDC#}`F=B$Z>WQE+|A zQL-l{)_y&jyJ_C|FgUTGUqc*Af75aY3{pq3zvh^8hv?na;#ZD7`rhTZX}9B_6XDTB zS4xq-CsU6W98CW~X5l_IQU|FXX@C!ZHKv87FMcU;$6#+M zo8FK#$Bn~JNSo{xD(NEJW)tPe=hRStS__FgV7l^ksjeV=Nk3TkE{ z3ck3i1M29lLz_>)8L^gOF=uENn_=^#F1t}csLrLnGaTPY!nlx3rQ>HG#LAl3!N5yS zQdY_Lc!fXnln;vJ@VK4c|+u4nwUJ#0fed_lj1h* zkB*UNk*sg{KTx{r$sP`e(;mSWG2+}^AO>CXgb^sv@MhLJ4 zN%-IN!)_<#2g@FBCXrwrnjdi8xcj*E{vP?!nYHYT0S~CS^56WVxKy=&85*KPSJ0Q} z%LxzDC8w5gzs!}9#&_gA$Va5@@>tIgW&JfedqPX3J-`5Q#zc0&IeAY9Qxv>$53zQi zS9=mI?@OTgpq4d4M>M@#wMfKvS4@2Td{AH7-GohqSp ze+YLq519-hQz9T+gkPA9rfN<&CvK%-$2c`5Q~vfOLRMtU>&o}FU}J{-F!72iW{Inq zhe7HOqz>pe9IC4j)K57Hza29QgE*##!Ot(j`VAT_`gT#bYBp&m!Ze&e!QDv$KMVOm z7occg?xRRK7!MJuOzw#>$>Z~KR|ti4MyZ^oGN8puY+Nfn*T?H$-jNgQcAKodNO~Mq z(3h6086LI3Z@=O4z>M2?Hft5kqnl7M6cQF5?N3ZvDX_2M$Xj_OsVwVdZZem;R<7;s>+72OR}&oX-}dea$WPv{WFGlcs< zT}&~MC2<2oNwffkEjpGx3a@+IPzf*V3GKNk6MhqSIlg*q>eCZ*$`g9N_xLBz3{-fi z1D>NPo)d=x1|;Z%r_659g(~uO&v}8M;hWG2<~V>8z#ge|f5G4*=p+qlNxy(OjJW`& z5awzcMzEw2cE@?YiLw!WCHKD9NpR>>e|vwtRVTT#AD^cz`Av9CZ+H?)e98(bg`79; zUnP4P+Xuai1FrW+LWYLw8PoURV)u8XIhx75uUrtHNQmz#ry&ds9?ZRq>+h3I=W|n@ zs9fIK<#a}be78|im+G{ydkppLb)G>=kmp5rz?Os(#hV$vZ=FFz`+ zr!){u3Vl(WejW?Qn`TAyAf28Ea8b~UV9@>+F`i$ar&He>LheXkPV7#wW#3=NO_Nk2W z#ZWru6|cg#ae*_c6f!^W*d6$|rv8YFi-0ORWkJ;Uz?wG^lMs~XZ4SP4zL(hQXP}B? zg>q--!2(sY;J$A6qnXmMg9Yv+xLCH_c2xm1Sh`1E2ZQPh#Nt&Cp;y-lFRY61ag|C4 zmi|KG`DZosO6qKAND1dpaIp2GD*NgrZN?>|#J&hAHx4xCbtY8L0lo1qm}dlm8x9c| z$&dF0_*_{kH)PF->kev_dtPJyn&9j1XFxm0-)-ZU(?jX^x(lzxoE6S#svQqim0XUj z{v=^^))uM{jms`Kvo{?iaFV`YEkbH@J~m~+GrJEauu4-E?Hi`)dkG*!!=Ca7fjoQE z@qU-&y#P^=7?j6>1; zXj4g3K5u})*RGSH01luPnS-L8m+cB0k)CrJCo*7f0V@5S2NuiRHk?Pw_o_gD_YD-} z$^wPofpDyt2d?ETRqOp=bEDdcO{+dBe#rLKKebv|^z4R~lmygT#E{32++iDTw&m*m zuICSe8i&FBW@D@xDRWag0$dOWG#xy9K$8)6!=*OEuyQmX0YkeSKVZ(+-$g<4+XrDg zCc_?4WMnEB5kO3HgJj?1C0CrqJX=|FfM(y9%f)R0;C>^t)BTu#a$uRj8R3hyDAx4620ojn7#^%9rt?U~!nW_pY?A7D5yWj%Xbr zE&3eagtredCGzpAJaN_~L%l&Aqu%S-qwf$*n!FyS#{D#XJnSjhw0Gd7SGXrDi=Yyw zbi9HW(J2}R>*VP4Fk2~?81L*~T;N*)JQE2lP#lf-`C`(~VD2@a(F&jMX@Vf8A{2WU z$(fKMANSuw@!`D9bVR$_uCU`AQ=;ka&|<GtqYd12gsnCS`wD_=V$A2G6C67Dnyp zq;vtdFN>{K@@VW;R>VAvgtDaar~&hApTP1+L-y;gxljNX-zwfGNfC#E12iy0^3nW(f`iP#|`}L*DuNmAT zOX@LYX59pDYgot334t6bX$K>L<$$&cT6#~Q$;Jf@C(KPP*FOv5?!Rn5y? z30Ogb*t}U_ujRN4uQ2lb+HwkeJI88g>hhv-^98VcG)81XZANVrePQxBoo#IFF^r{K}Ha?Rdx^yxc)+;*aRn+g94hsGs{|);?@B(cSNd>2{KRq4$xLH59bgIu!ymCqN%XV7p@DO37#WtSAb? z0TaIo#M;B|Cl0lhB;BwT9tzD5&L5y)%s_& z^=IvaoOn}s^2axr>~q&cmm}K3!c!mL=K55By~p=kh5>ZKo#q=ee~SH)G+-i~(K*zz z8NN?;DSQs>q6(7RK3l56Buhq^)M4;^p#H%R*Xj*j@?V{4kd}Ogvh}sW;|C_$X|$_a zLMX)*`pYxN{w(>)t6b?pte~ysdE<1d7#$)TU;8_O_vr&m43w`$>$cklT0cRr8=5hB zPiuO;CIjGG7t&|;8rO!*kyHC#-73c575W~^F_`JMkDE=&1un&>7W($$nK2k>SArGX z6gP1ILOEu=hvhP-JGpL`qmoA6)nzmdPWSujq|)VguFim%zV{S~+jaxqK1!M{i&$st zpi91!6PVTSX`9Y34nFr{Wma43co}Gl{b?D-dx(Q&%#_fu|Ck?@+&Rm^r#)CFlo~F6$ws zsa1Q^zkhvj5uyp@U1vde<~~je6U`7up_(O_``TX<&mzt*^T|ZeSGV!4KLD{s-WdSU zRNODFkc=LY<@(6h-|!37BZM&pQwYqv9t2~5&R{X%ZB=J^&6=#jM0LEP_7f2PI3-r8 zTughya-#++Kc2}i)Mq+S`>)qs&{KLR;V)X-H98R%pB0_ ze*;8(%JY-B$AI@_cz2;{?!nBh7L0KXWD)rB%p@7G*HA|}IeImhU`o&9RC>}a4TCep zHvS)fh?<~DF$XcVeFdVIn>-)8h(c}Y53hNl3sr%6kY#(Ts$|S{~teLvrgo7Jr_4y7SNSZvSTiczMVu(il0U6JN z#vWk#^hptUDiw{h9`D*F`-)NxJ3v+_vubu@{oxG2| z^e!(glnSjw5U!+Pj*@F=9E8v0Ib;q*cD(4GB8d(?cvN$oP6Ahz6(etrOB6m-aHl zm+%nm;mfAz8h35(zRPFaE})ktKxJiB_ws3o=D`DL<&s{$r~Df-UHeoe=D10rJdILNeIk_-y0%1mI+1lNNe=LwhHLv_9dpwGgV z&*e0Bh@c0JJsmNDE3at5%KBF;9To6mD^x#CcZ7gb(3kx=QlQBw_JBA5*%nG@*%wz+ z@K^q+LWGEF1c!?{Olp1cF?RsW&T+SAH9bAarCY6CVCo(K3Qm7IOtFxsNaEL9_X$F@ zaAJfR-Nq^G&*5xMx`v0z>Op&f%ICgvAK>=|HH~k+*{_O}zvJ!loktig^8~$$0u+IF z>uE5hzGad&y?t9*fDoG&wJQhoFvQCjfMidxtkKWwAzh>f;?(=}7TO2J>;&EXVnAMi zMK)wQ{KgpfDh_|5V@c?0k4sjyL%~5v1~7~)5(`6qiYq5b{9xG+TSzZAUi<*w$^C*p zDL+Af&Y#2>^wH4ydg;aomCY>2){6`Cf&zIQL{jDG{qvJ2FMnAP?Pb#e75|=nL0-vU zyl%1X_-kVPf$Ck0AhZ<&O1#G!i`*$>E}-> z+%MFb_GI)lHK;1$NUb3Q8IY?YLp$n~N74*-N&<6QFWTnM?K~YY{0w{Ei2}YwBvUSE zt)V$E)aE^h0Zg#+VXC0SE7YIJj(7)&vL@#($NPq!01ULjX4X1-Jd>tM+w3x#1 zHNklWlo_1DzRlqzEsTe)k9yFj%dRoE-@T{O1EgLZpCcn5X#r3S=c8twy1il}Zx~<; z=<%3CmrA?l@%@yH+R=e-z7ET7M~AXI;H@jn8?K)5{bIRuW(FB5ZXcX*GhAIEjhDhf zo>+NmD5J;sp$E#lGzz;lAfHC^hvJ*_Ese~qn4F)zf#_4X?fK9RevCtmSr2nr_$HrH z;gA7+VzxBbGkI={8^LzxDwNPX+PA0kMJ|N8KB@Xa)u5{2elJ?54A*G+gld9gNsy=e z1N-zP`*@zbV}l7ic{jFa6GJ}?!v*p>K0NYz{Ll$qrO{6=aN69FO55(jXKFbG2+4G| zXlN)YQ1BlWl@V(SO}uS*dA|<}$-=CJmL4{~1C6u=ANtBE*p++!YFV$5Sf^8|SG=U9 z^yMslp7|%6mo#Wa015ekQst=p&Jf%}8=)Is*c}Xd%jyJAsl@8QHt#V3z#>8STJ=R{ zN8$)zgLVtWvpOX_%mpT}&Cb{EsFUci3g_sCYbbXpKf+2@9?w!n4&?5a%SHyj_!Zs3IafUP}s+>o`*;R~($N2k7_Q^Or6^7~kRpjOhyT7T^mH49Q&OKPf| zYH+aB5JJ3`?}R|^N3Q+Xfu}KvlJG?XKZ>MEx&Zv6vS@x&d2?^>WOAy4N(a641gy#X z-au#cMBS+Rz4l#9o569@XIb^gr3hX)r7I+)864bCus|iz=bIiFjwplLi@WaOrzF-B z-?OH93={}r;FaTA&Q)F#zBOrkbOLz*4$bf1F(0!J+`S;3O1b?@QbNmXEAjW0!Tt0k7ihR3Yi4YM+fu~ZXRr3= zctLsuWqna|N~bbZh^U0O?W&)+!Gc)C9ygL$i>sxfDiu_ODpuJpsJg^{lEJt@Jf(~u z03^dS3vd95YxeB!3SgW_Rsg4{sW!kl1~^X9z9zxmhmjx>l;cfw`&6%N*)mYqY~-(e zWC%n85e?GzP+s)pvrhmsZ?LA9x=2Uq6XqA6wKT#-(iW8gyGJD;5AZ9Cu zfZyFuq+HZ$pS{^IFFuGxp#yyo__u`6!})?cY?}{`eAmPRd`6IJB2Tb&zi-lJzTofg zihH9Hc%E~cN4b@4L@kY!M)i&X zf72*ZJUb2r_2g^4Gax171VmWLRet%@gCT4^G9d2jjGwT|8h}`j!oF$l zciqZ__XPb1^l~mzpruiwU!@{a4N|cc@@siPuvFd0ArC*x+;7KT$_EfO*&i$|FHZZ4 z)Gb~PTcYsA>yK?sD6j7csY`NQL?ljGsf&=cNBYJvw+nogp-%c!bf2bFr*8tm}9r8`z zF)P5XBPlfo!KfMW);sohtRxm6dj^g_1J>Cn_^N(SFkR>r5Aua10o^Dz) z;o1H5_jI8jUFm9@Ocijh(e)b@Bn#DrLK**k9yISIQt$Kg^VWavR0X*YUQL6K`hy=@ zQGf6Ky`?^AZn+{a-%3*KV8;?x)ZyORi}6x?y>a0tV?l%7Vs~kW)A<=O7j3>ZaFQ!u zeN8|nmJYLG+C!WY*dw!~iM6e;l6mXwGMvM!{rCb_ky)C^YFIS5Os1P%=3un$<_f9Y zRTAc9gCCzogPT#b<>C2y_2qmsK&5;3l43xsF#7cId*6l!`rt`;yXGOPm>qj6L4_U__1`(R4L@mxS70jDKBD7g%R7mW~KFcD3kYnvHlc^n?sry^^n-F%<_%T zQ;$pC!1^2*e9*p^PQ{c`cY%6y;w<|joWIW>c^E4Ae&g#-*P&(AO|4}OEQ_M*F1u7U z_ofzQ?`$Hsz;^1e$f0EZlcZ!riG;sZBV>Xf-I0p zo7gEZVdGFRhxsKr&a1L)is8)Ug%@iFKuR&zqgzz6hpS9~=a&TM`Kt&xfVv^&W`y72 zem*RL-4n)&gZkE1tbhTIfCbfqJrRda%^)zP+^)ALLHj<}&R(74{IgYB3f zOdA^E6daBAtpsS|XEVBx{ZrVj`6T0AKJ~Bi7_vRCuYBL2FyWhk1r8aA6elWR2MbBd zUN>OnN-r4dVd|Td53Q08>=sisusgork`~bIP6Wmioc%!UdKl&Qjn^ zy5^3l*O0$(2G4g2`KD#$?fCJVGo8Tvr%fT?ay%~OpI~D9BQ0ydsONLXaO!w|MAqv> zO&7ZfeOpw`E6>N#sYc7;8XfU@lMbN<@4%u31-Xk5x@;5>9x!rT1$c-;uvnbxzma%M z0O=4XVB|IK`LVP&AE*pWFB>^pkrZPwNItj4FU-JiUgYsh1ZjT7K7L;8ML-JObg7yh zyg|kTDw*;RZWqty5rf=7iI81@cUgy7?A`>x@;Q)~bWa;q+I`4Jx%11}$-%d-;oxGm zX0vUf**)L;p{NCV@7aP`(ZD*9bz2UkLlG+g)6YIqi_%Js($JmzNq$!c44=dRwDVi4 zm!ivd9uTpwW9+?>YP&L>9EdY9R;{KBN?AJ4RA6&Rtr90newCiM222~G0|2i-^r`o| zovW9I&29k%9cEugpW*a{D58hS{X)U?VI0UeY69yp`!OVBsOfq2+(5Go!gp?|#+J?k zYgxq*&0~C|$W;aL0nDK9FMAsUb7bc5ViuPOGxVIGU)nj4f5-0}NBCr2!F&zixJhx| z7*9ee*@zh!14;ClWj+WJ@j<&xSw9%)7}PKb@=^+pOn0k< z?(2hfi|g5H_h+Awqx@VJk@~Dp!vwX~l_KXXaF$t;60Cq5bmOG+)0S>L2+1! zo-098Nr&WSZ$LzS(4y=3O+_N@t{?GxD4c?NN5c$-hFJv>qF1HtXG( z^46j*a7G8BDvGdy>VClCX*W&3Lf1z3D{qB9{!u_&oX1pGQMX&k%cxRY z|3FC7^9FSso23K%%ns&b<-0=}qN!$NIqZ1$48ap8_Qxqw0wZ8i4UfNLku8#NL`DI^ z{YiX5P9A0(uS-B}rAUCECOMUhrqA~6H{)zYSvCr;qzJ*%1(qUc+E(w8;lh5|Lw65J zR?R3PtM_u{S9BGu=t$?XLj3J>P}3%VyqDgI(3Sb^+o)?)$E#kGVU-0?SPe?%r8cn4 z>?5Sf@r2<$M~n?z$pD1#JO>yb?m$Wpf!T!O#Zmn2X-F?h*U7#ufRs~eP@+g&|AeM{72GxZ7B*xLG+gd!SG1VAUV7tDp^24 zkguPjXRTQidb({};ofu4soI5**1n`1u~4}A=hr4Qb&QO`?jvKiRf4GeC+I(^%uI2v9F!!UxP7&UAE2x_mr)-gV7@7()VSK&p#^dgXb7;5X0q0zxUVkRqL z>%eiVRDMOU>}MPD_KhpVKjHj|r90X`v`5CdDk$yync~dN-;qBojh8tuu}6ELlyl)m zHIzGq2+HAd&e@L=h={|N4i5nOtY&;!(JEeU-WQr+7a&OJ$1+G{38()2ebl)InZNz#iYrw(?igvJZ{+K8vx=+VUQ7 z&rT)w+kxWuXmW&;z%w&vPZv(BH%qn|SpSh181Z}J7+gfEd5)wvhNh;k@JjdS!H?Sz zuy~dvNBoqtAl4ukVa#>&)_qK&t0uyn&*&_|UrQO>TmO3c!T#l3EtvxQ?2?UYzEB1p zzASnkkUW+jg|DTbt>L8=`P01_d2P^{Ks_vexgYQs{?nn_*qpA+PCuQ#=g+t9y%9s5 ztbm4G$F|&Hg}{=4{0-Wx89%z(7N`(XUjga^qCy57(XflISs*%w`@u&`7l7ymZmH)3 zxU{8w9(MiYSL@Ihfs+;;7dFG_lG1yzjLRmqsYe0bjG4ZoZ9Jk#(igPlYjoC{$M5z4 zi;KAn*KJ$$BjSXy2kFI#UXWe78Vg^^-(}|$Sxm$rc$9+B^_<4>LB#*I0gb0% zBvSr1okA)X`hh(fSwEzF9wKG0?keu6J5M5PfbcDxt%f@Nge)m*Iq^)7(+>B%j920Z z(dGnEr`-8OgA5h-tyQ-0eeDJH@rKUqAY2~lb^o5lfrNjZapKHoheG<1ef`M^9>-fL?6-R9t?-eybv{`(8gC z!r)aL^7F60x_NOJulhvr(;e%~UhcU8>3DqU7sDN07aoY``!rrJ$uCme(Z*%)`_}$l zc*@#z<48Wsrv)j`cf&bG#Oz$3Q)<3AdN>&Cyf65?y>uy~q!q^HoG2o>`h3|aZ%v}B z$6@Az3EV?vO@4%C_EqEgi+E|U&!0aqMBmOh5m$$Ar@h>q4Y{w;>q4goJX)#53FiH; zUY^Xi+>D$~dB_2FW9n|G@Q~R)UZQ?)wP4OIdUv$F<@P2dVg_)GTjpq1fPiGe9e=9D zl_24HkQ#fEGX85Xv}lJnir?87e}2STF2*k(Xj#m7f7zpT%_N{{_9x5yq)5Mgv@a7@ z-a3AoT&WZs@N*9DX9xF-hD0Wc!hdeaIP~fhW;R( zI>hGse;ERQt47*3S8y}dMrkNPlNR*2Wn&yMry!RvmTWe+IC9fTpb22u9|N#WdUT`q zJC@pSH=>&E%*q4_(%a_#&BzFwo$|t5xtXDZR#B5;fZZoG%MV|)iBCF~^KG+fF@4M~lh*&8g~^%d<_!($Ne zsGun6J(ZXG9-%j%{1ZP;kXdtr52?i$3Lx_Y**wql2vNB1b7{!e2wA=E;5?DS{lF@- zfh5w`l7blLpKB<%P^nJ{raB%Lr@|>#z-O*s-aX-y8RYq)P?EkU4W`q1HM2T`!`-bV z3AS+}e?Q4%{y0CN84?`Gpi+YAUfvNufw`~J3lH2B4iZ;n->runnY_@~1mKvbcZ#{7 zoI;L-bwN<@lk&_TwKwQemD3+y|DoWm+po7+cc0{mm7g}6@GhI(?sp(JE|=R=m~p09 zxySC8`l5t_tKEiQ&X2P2?)Kj=jb=!BfM8Hyz8&*;8T+dtBaHETl=K2E`|K33?HT+K zBxvEwmEZQBFX*kT8ssVdNW3!>yFFnTAyXs}oXv_+9uKBPe|%y;wxai_C=3oRGCSl_ zBU=`4#uK_t0&itGz8jlhTY&Sf;{ISd_`sjp+Wh1vvziNr<^R5~eGV5Sio{f%glu+~ zks@JN*NJ>~>(ZSDVHf@t&T{U>HKL2KBiBWu2E7>YfTuV`_-S$-T})9o% zJX$K6;KUdhs-Nt=Zd zE@TuUvg*dYx^vd+C6)Yl-y+!B9q2@U_@<5@K!nvFwAzRkotl0~cVyf%tnl2|zXi!H z+JX}9o>Gey#NF#`N&QixI1G#AJSrD5BO^# z(*8__!{afW3n1zWqyO<$RUS!Sb^QvT#uz4MnoNJ4_rA{iTqIlNxOQbbp&_W?g)y4! zVW9u%kP_jZFp;Tmmudf|ACkDeiAT^gyt-eK6o~3C=5?&z^lQFfg22x5%&C<9MZnm_ z_3$UMgrn?zf3|ZD04=^qK@!=>I1jDPEAL22ZN7Hx@kL_*V?+nkTkNaoZje@r7TCOa zL$Kn3$cgq8`zB~&gJwtpHA=s>CnR9H939@m3@2PuCP z2nZkabG?1`LpJx@nyTS@$*1ZPvs3LixPPgx3-9?IUcC(q^u17fzgR%RwV|nki5mD- z>j&B%VBfC2!WP&+duC0IY!i?XkYch_7YTH1l3w{Elna+Z+dwD^Y>M0SLE;kmlpkKA zrD=UV1_CYRK%5-^l@MLrHe;SI$C5MR&+S-?5P*hsx>xJjBR_ozgk8A>Ds7vyfn?du zL+pWvy&;24mVuM7%!@?x;cXD}W&@dJml01>cTe2*$YLKtE>Z7hgGl^(2w5Y}Iq-Pv z)#unCnMFVBJZUp2RWGjS{5_LdeHhO-XoKdDkXfNSE3zi;Oz-|zHv!|q0pYdcQQ0Pc zP_PTuRJ~M)q!fj+2oY48WBXjx(UtCCUHY=~yDc9XjGa}%WGfOFD}Ef9=yAfa=cp&; zcA}`Q(E%64{Uzo<`lS{)DT)4xeL;HS30Lzv7Z5Nz?AAE+fn{P+ePIGm2;W5t@|7B& zMh?fFe02S?=MlU8Ms`=VH~)$+r`&k;AxOe|MvXtXFofgdj4*}E5V8SOoqrec?PWs| zczJyIi{-Ap8*baTehHCvr{?+@JC=a+eW@DGc#N6aVZ-U|9?cKexVNLc?`s!^9G?h9 zKDcC{xgqC|UXvV=7nk*)@H#ms-uz{nsSUYk25By!y?Pt{U$!ooK|EY~a$vON7{32W zWPf#uy(1uBKI`}`&okBUcpH|Umgk+fhW1N>WQjc)3>W3(!!XZv<>`(cMZadr!9Bd$ z%HfT9RN1>1{P-Thyc9Io5a{K?YwB$*$enh2W&0)q_L7u@_F%ECEptJgT4DORYQUKQ zX;F`8s$e-BN#lkR;AX+|{#bHKytS$W3cjfa1N)okF(JyTu@&9 z;A6x74*I|f#fGRa!!b8LQjgxKKV!akIR(tzYv@GPvoSqnkxgo(3xrpTm z4sql;IXdM{b(lMobiX-2mJd0o~b`WfZY7{W8%mzM5L%#p&<0?E${ z&^Ep9Cn;T_+J%z-B-|fXfbM4gCygymdm~=96v=yMpYej4cnK|RQXt}fLi%>7*!Fp_ z{IY_-7rCMM8IG3}N#!r_(~?3|Jc-dLoL7HAN&|E1b!2PNm;JP)K_~S3jUW7VKR;vi z&_1=xc#lr|KVK-^K0*khx{iM_)`m~RU!E`k`de?~1y$4jv!PJ^B6--&o=zJdxWvZs zrqB{>)*|1Z2lKdw%#YEv^awzOEPsUqkwT`|AQJp=eM4WWRvN;iI^Vs@_(m+8JpG`> z<0krn<}aGlc7LAkilKPR4EV-9piz;D}TOe1a~%|Q}Q_tr{E+}ja>t7F!=TErvR zz7lwt5S810-B-c2M8U5GW5IPWdKY>rg0tc6HPs4jr2Ik>O;ar3>xY@}?#!PV?K;y_ zqoWdX(E;QYe+(>gst!}>Dwrr0J(Q<}L`Mm*3a0%=Qq92`+?WAmAYa9t26!5{t$m;W zpob6`;R8bd)4sqNvBFOn%ck&i;oyNJdwh^l9@FU*SG(9j!jCOdKkdwe?RPvW?eS1^ zqP~qfU^6$+R31O%@V%=FS!oxf&DVJi0nBn#)dyG;QQg*caYT3W4_%FJ$I~=fkNo+o zV3J5ORI{&h_Q~(CAiu-1^MZm}SiDzKgC3ksO}2D9=P7Xthn|!e#Uy~oX~SF2J~XdUET0eNX8S46#>;hdP5^y?HWGk~qVt>poG-YL zSBZxs+_MtDk=(?&_DW^h*DiANtwF&3`;1;*?hd-!Gx=*l+0YvzKA*C;pqWZCwmrdDcr> zUFkkj=0=UQ7Pt@K<(7_oy~)eVLNcBD+gDz~wzuo?YZCWUeV@{XK)ifLT0SDHvV9+Y zvCt_SPiB@T1+7+BL|Xt~_RTrSIwL6skk0u)?CzfSaKFsONTGwTP|={TnZC1W-gqH< z8iQ4TbQr5$q&P{{)4l@5IiB~CN2i`{8u$uj!%6|})bBu2SEn=~Hm%D70yVX#UhP~F zXP9g!Sx#M9R}e+bbxf}Vhxs}oof#;lWp^PK&H?eMPA=sDUxsJMJ`uxlL!*$--4Q9= zyH1#%y@CEjLQTlmALZQ1r+>kb6uS}pS8nl=W*B6;o&_S;A8B$M24{<-TO}r1F?DNi)!5lD$K`y-RfN2j|BhdxYl(Ne*RJg4-u}6uoM#6snL#eyr8l zDJ~n|yo^f7(FC~hVYa_|`oFZYNcC2V!9IbH=NL-nD-3}wMLa_oou1y=Ra+0K=?@^k zXsHGJv?Oh%U}xm~e5$t=h(t=JEvOatTR#1qyN5l6I$x9vM=kXL>_Z#26bkMYL6R$H zb;AyKN8s;%NZI9~3(tMIq>s|fXyXf_*?YeF3Qkmq7J}L=@)kB1oDiNii|o^5_PWP1 zLBbZZLBeh`C!l`X9$^!IZLA1ClGwZe`$l#)9^Xr^!4ZGLvRIXKW2qNn2UUs`b{bYe zV6%bV2>u-{bzN)Nsz0Ab^qdlY&rB4XZk!}og>9mP857oV`%xhgaJ?2oF$S!yWZ2aA z+0>V|hJ5FViba$kW$=41c8%6JG(TadA-N+5{7kCS?9g!W%9s1+evTTA=IBBpa=ds z+fRg&rN8((-tSmsIJ)9X6?73lBwOY$FcPbSycgKQOS-&6Yu}=#y&-Zg!Sdy=$QeNk z)gkxqllz)V=%hSLG9zeetOk} z^a!*D5_I41B){=!)rUnID1Np;>UstsxE~1Tq$xkDysRa_$QHtd&&}%0%>b%$?j!{V z1i9u)=H3zsiBp{xE%~jC-))r@&zSh!E906oq^H?#a2db=dJ9gm>Ed*S0auJBxm$#2rp&n&F;wph`iy!5JSf++_0`u8EM4AOmjz-C43#H3fp z7r+mSkYWeQ9Rr?`YWqiiJdOzAi)ULI7p_3J8R?zNaG>S87jbH`7k`z*>DMCnGE-Gm z7Y+7GV>jBrZ#w9rc!mVB4s4$=Qz74%+lJdO{OdjL77SJqDew6RjqUyY#nw;rn#E|T z4P(ygG|FYH^}KUt*UW(64}F4)B4^wiVm{&k*DJye-TI>~y*6a^L_x`2g{SJ7f^T2L z&e^8Z{<5MuFBldBY;Xv9=c@6a0h2H!4%MdIV*#dCx@`wzW(h00&;7%^~&wq1% zpB|F-LDbBRFp(A(tZ?doozIHn7>y?wx<$AAG$nD6ZfSI{D*>hd$wDs@2uclTPU)p# zdc)vkJ~HkZo-PEx^~owzB$c7T)uwKM4t_p}kI>8s*VNv2^4zOE)2Gq&Rl~{>hv1jO zPh_6>*1&SzKCTX)J^PT|3Y^h+uVrAEppYq^oIn*U#;B+aY^CXBJIeTs zZ3Pk5$o_+f6+Dt$42G4V-=FwPS+G%3&!kgRWU4~akpqH0xXi~n^32F?58I+C^`gS- z8D~cbbOA{N;o6S0d__VS*B|vkUBfU%>U~325x-?whh9JnB-cg2m1~kNM0}j|Lq%U1A}(BzGJHEc|ZcGRLF` zAO-l__RxEGsR5x|3iR7+UH-xr{9z_ES?Ol zLP(3}DS~Nx9)laTKj)_)L5G9C#07}(r3PGy;%>ZeT(eJgCoI!)BT28D2Zta+^#VX`Vp||u;+}n&0@)E_D8ZZ^L7*6(yDPdg{ThLiMP;*gK|SrOkF{~dQTj{U zV55pyO}R;4g;Bd*huOBQ~i&a)+;x1Z5QEf6ESm=(w&PnS@)8RzkX5h2tSWVC5Q9K_TeagDu5Prg0~9CN0}K| z8Myp%K-vYf!Ci*{Mf!0}smB|PWPPL;JQa=B@s8ZZtoent=lOt0x*U!KIvmRD(cMz> z0&&RW99T~v;c164YZaN^b_+ z?S%Fc>lnW112sV7ZEY1k`eNVXHbRYi#Xf;_@57<;9qhQq_t}tj2^C|(Y4vf?{bx?| zPQ$A3Js#A&W>lY{fr}@;6vS!E>SA#14<4&9kgqVS7RLVXfcATSd*VZO4ULS2Sp*o= zE(1~b&=7$n%&v6mlupveky4@I0)bA)R~tn45E)`-d8yM(&A|$H zGvNLiy;3#*I~6GMNIkP}C}PbXNhwfCSYND<7p> z^xCUo%$3Ul6V ztP6e$@%)8RGK2MXmN;8%2<=8gKejIKeK8cT#lD~5*UgckUcW1?l;kvd|FbCQ&@uQ3 z$2lEp(8UmyFyIrfnUb!{x%ur4juXDYOBhXev|M8;QlUdAIBc&I=uL--Ce$%Ag@bfU zKpk}jy2NDP0>cNVFX|K^0WpA<59(WX`n~ou=)zv27`)GYVMeMIjGaDlL7ioNThUL{ z#of0@Y@}RyrDn!8wj_KW-{;RuSSrzAML0r%%<9n`gMAIz|yCCZ1ESt>tz~b5_Uj9 z{MQlPMjs|N;NrtDI*66T1Z(*NPVCiGJ`HGV!xg?NDi;@^4C90G-8`_@oC0nF(d(Vh z2A3}5$QAx~YV2!5jovv>GN88-XRJ=px;Mny{+9YXRQu0q{YHcOo|5Z}GF8ocoqEP! z9Lci{7ikXHhjf9V0O#OFN3ie-V|{;)rv1l3Rs+7h(W9M*ZZlap#Xw$NU_&B;d`Wrk zKvU9X>wSJd;LbYnE}>f%0Pd3eAs?=!u4zx^V<~(m=3_KxM0l)!){gLlS3?WJ9@qc! zI3Y~~9r<#rf9iw(_+eZh-Sp;dA3*!2KFWO&@p?%7h9pBy&gTn>x#sdgdh9*p@3$Mj zw-vjKOMNcp^}QnfkXQ4*x?Z(QBi2N6#)1SP3xCmHm*5u(gnK!!VW1^Qn>od*9is=H z-p1Q!+pmg1LFt{@aRz#BS4xuIYe5fa<>h(3xSK)qg{t{iB9?z`egbIt%2q!&RMk`J zELp4FfR;osvgR;|N3j(S%68mOeplp->cZW+uRZw14zp%c>RC+>@Y2G#e?Eo#Jz6ox z63QSwIC>6Hq#X;L^j@b#p6VG#2o|*|{dsYc<70cJ?s@OZv=E$lE1Sdp#IGj_5g=OA zb10d)2bbdW9K_hGiC3Ls^YtHjWxrKW^L+2X$!3w+g?nJP`0$t18Q{xsCW{{;Eb)?C88F_)_tW3$dvXq# z(|_0d6tIaF=|-)^atF_k$A;!AvM$eafrM8R{w&U} z_k358u(W7Q3d%qXA-&%XuW!K9fhT^MgCGuBLg52v*9x>6IP4^vJz<-5w_Hli+(Gd+MqswOk6>P&ioAQQmE}8$fVD)i-fDCfI ziSuFmAr}>oLPdV>_HQaHd&Q((BaE>k55xM3kvB2Om+C!>+nZNGSvYGBof>Bvx`({J zW8RedVD8be{*6&Q3Kh*X0 z7o4mo$>P>?biTc=H!w$GG|(sOdx1cca6x(k2@+59B^$Oa9HDW37|-#whtGnYB+Me+ zSNT$Tmof5V);+iqv+Z}xH)|I8k8~4&Zbu=zF%5V4JRappuP3~rMI9Z!_{{eiij%{Y zN8@!^kQYRZX16@wXo6hvEYb`+kuS`5CV(NLE_~u`;ZJMo;ya(2`NZUcQeD z=LLEQfW%*-&s4~pfVNcgxsJgFY-8s!DjG$vp+*HOb{yZc*Xr}_KVT%a${`3-RrJU5 z`}E!Qf_9~)0ke$ZGMJqGF4DvF%>O$3-2??53B6hPS})n&q7yRxZDDqxd37wfrWC^6 z8mm}7z;1d&2WoSX9w9y^#OK~JOl``F!n7YLe8we3R6x0HiK-fxHrAui?|ba1as_nCs#Sbwe|};X6SM|KV$Koj8AU= z;k}#wdjr>$?uAc%WiGl0Ez7sJJxW1pnWa>|`}s0ZeRb2}^XVzUUjf`lm=(?f7b?GG zr3>t6qoSd5`wbXSJkJbCpT()Qv;~u8>-J{ssbn~;2KW}w&-G0?O#!Wd{8y|k;Qnr3 z`TbHqqcK3fSGJiYp%N9pN2t~*@vY@KeTTbo-M_v7unC_emWXoiL$EpTCFrq`_j=NU zD%LOJbGn9pxF~EEVVn>M^LrHAi4VWd<-Ij|*>=aI-uEw6*WuGC?Z)Ua8{G@-A$9{y zRrXgR)!xYKY;R>N3fPQ^e6j}*d9~7_+~!KOIM0W*G;sVvwJ{W~1kBb!(Z9SfdP8af zruh$uzNvQ!-jc4zYT7L3DO>0lAP!b?6XCdMLu__~yZRry?Bi3D4?U?4*s}5M8MZJK z1aaf9{UUx|-y~p{<>gn&bWE24KjX#mS#-{thYnS3DVfH{KO^@thz-b3<>pnZpl652 z`weXZk*9{I7*;4KsYxhd7etTKVYrgdFnv4CmVTwE+csGWcTpEdckAX?D@e`=>K7@HQHuUjg@ zqC14rh5hv}?8+>&Z_UOA4sDZ`DV(4yBc@O#KX3;-wi$i9@)79ny1~(FUB(ZKJTZVy z`FBXjnPxpGiNI8#kD~?JyaplShm6Avs$9?&V~j_xpA78qY)|r}Z?~cQ(SX~APeh98 z=?7lRP!D`9md!=e=D0GUrsA0<VhV2LMybSbQ>Xf5rX4FETK{EqhJw&D#7d$5D^w>btiWbnQ=kSyKK&Y#8 zsLxJv7bjopuWhUQd-1Ro0_Mv)bgi}WfsF;T;6OU(bmar+g;hg*`Vb(eqK!ibgzcJw zsRH`oj12bQZ%I7=Z)|u%9;<;RnRuLMVSH|=Y{murn!mHj5efCkZfb|f5$en6xZPb% z(D@CZapfvie0(fmr7Cc}DRO_O31vGDYlFanT_rsuWS~_WI;Dl6XJ%&eycRd z%W~o_wcrsWSHWbSr{Rgbh7)|qiezS0Ym}koQZ!fUv!fV-+tK4{Tbds837g2^s4hsNiu)*p+6*x z<1<{CZ2FnE7r7k95SO=0%&n-h=T-C&skz)aw7y0P23lgF!nq`c7G`%y1BQ&(v-lKM z8k%l|b}s+MUgQ@-F2`MFr7;;0%%$lD@iPiKzd=hY*#{|DyjYw~NUeXuG?YcZvyRb! zSQ#Aj-o{j}x7)Ta*9>2dV*>DBfWYO$0SCQzDRS{T{+S`;{rUj~jx5k1^109kK#HLX zC4EAd^<%VJYElCcPDqCf5}ZP~)m-D}`FVC?%-v=2sDmH7#do)b#;4SAygw1h${^Py zS=Fv?i7|cUbG)Emr1}EtR$|%S9vOPw`E7*y%|6~S(DsqO<%kNFMTz0yP2uFjqqtte zLVCQM)bxPvhU0wzxu(u|;K4j989W{km*I{Pp+8O)guIXpy!3rYy z(n1-Kt5#q*=?!ii-05OTd~c8?%fLS3UqkuW0i)pVVHdbV1U4!e4Zic|_`TfC;Vj|o zBlc-~^VM5*_?Wm)VUC5>MlO!rj@EH634-0gQixPjhgsur+%;+Jo%C`!v4NRKfEqX1 z9L`^LfNSAPQ=dgRwC&y=UzorbO~0}w^(fwts%gCs>plb4SI}^R0u$pc7EZxLTt_x9 z#I}D;ueTHhP1H|xNFKmc7n9P#b`Y0T`aQ}P)g$v)%f5Vf$iM}CuMm8Lgk1K}oAKgM zIoB6`m!fJ*f3wn)j}V!$@&{Q-mHEPnUMw$NWyIqomz$=O{Ae$Ie?OR6{EM}McRQqj z8k~RNPdUF`uBD;90g(doE_h-cW>x4oZU3?A@>l!75&fJ_fA;?^dJ$0NDa(CpKo9~Q?BP7r(~!!qHw_F| z3HSDEPzt|4umDV9BXqK4IGNY6M(co3fedGoSZyXmYOSO3ZDX^_u6QTeFd5DW*|9z;T&}PM(gY+ z_RrOn5hF&eqdxAUr?W=|>Bn$32!qW}=?RwE0%!#HlBROuqk6o}Ta!s&(+-T}dyR|& zI64T&IDZD=XK}h$%RJU7E|eeLQ!TE4zDh}15vbn#xP_aMC(!K=H}_*wx2&GwR%jfh zBzigP?sZMX7|W6ea=B3hq$CLR;(lVz(ZKVp4DuIX%(Fn~Hpxsuci-&^q*I5J+_4c> z)Re>do2aYwT4Ec=AMH#%5HOk94(ESnEXhDD^xZx>Vq0^+mx=MSFBVCnPnYAI-w&sW z#)lOTJ}|fU9XlKw)z>7P5bvn1CLG69vNmlEu}*;kHAW4F?K)mw_%v~$ z5-Y=`5}$fuob9G9iEL)F%cZpGmWb}FZy1{%yFH7PeO9BCu&wy>1DLfrDp~CPwtD6t zqEoz86avt#EmU!Y=v=S(0>tSc7IMO3@?p0K(enRO8dhYW?LQjtU;L->IgkFu7>4IT zq-sj}iT33q9A3_+IOD;Yzw)Lv*vnpAUl=)JZ>47%jxUAp;I@Cp&0i!Je&b06(f$f- z+Trw5xb+B0RmVqY&HcGAsscf~p(~oRDM73nOuM`#)}b=Gyy>z9k@SokO)NTsNN;UK z?frRf-tR;1un3!r1-xl$PcVSSO@je8O>(+G>T_$=^AvuN0(5<#Ce*xfoTC=HK-dR6 ztV|#>5ItIvce%Q)A3WR|0Tey&*sO9BMp~nTW5l5@75$?EBlMr8@D{xITSeY}()j#~ zEJbU$*cTyFlc;la-G0zj?1z^KEl>5*#YZwFF}@-?718ECeZZ%Z?`xbyx?QK(!Nfd! zqFdBoIUiEl{YZ93OD{(LFXlB7Vz=+ua4u?WIZsjvv-3AjrvkbK!5}0P0ui)Ag!grW zStZO2>JuF!0UTNB%Gf$ZU6!X!lIzMOcwhvzB6(vA+2=ACcCisKwV+j7SH^iiL(?@@ zCrdh6e@(E?#POs?6DU1vCWfz_a#5(~x5UaH+IG#XO^?cOFWSAmFeR$v{@1p4;a!n` zt6_x=mBye(oPp>+9f8v!H+;r^IWN3;n7F+OBZBz_5QM^Q{FpYycl4^K(`}v+N*(H> zvaypvHe^*-@AS9j9|$<|DH+6?ev0d>k4^MOU6M<0D0Y0Jj35R<&S_!m!<^OivbHy0 zegi&%!_B+Seg*7?;cQICVwf)6A|9z1gSv|o8uU5DTJPYGI9YHg_nj2 zRta~2vmegxd8>JY_3x%TJcUbec1|oskR@U4>0fLgphw*3063wne^lMVr&wbVc#MQ{q#=B`>~9zeCVynU|aPuaP+iok$ye`Cmz?f1#*Sp zzFH?}PS%3QIvo;7&F8b!evbleWI%g0Cleh(U};;m1uXf39FgYA_WNN5znCwOvV?c) zN(_9$(EdpbV&%OI0vQ;a-qcIXAbUVxcHVxoEBeeR0Z;fi97l+_0bhRuRZ?nbPXjHb z_Kz)U$zt*{P+a%RAnsN?G3O@gAW=tWK_KJ7#sqj1ZodNybvr*kp4XIqls3IU{#3zj zud!^2nGlec$V2(9&pdf|e2{be=PrB&9$|6lgsu)~4XEipK~%t(Swl2{jB(`G_8`_A zRxk@Y2&PyD_>M6x2|LQ@@@;hY*JZnx=?_Lpy*t3n>(B*cWh5k~WTGnB*ADwXRgIa$ zE?RQ4ZsY)2kw1Epr8z%$$v|dIWtJ6RbzlLqu?;-J1h9baL?invh!tv+GfCaHjS== zJ71oK7D;3ge?1j$q|cbSJZx6+*)(1R19VFetZO8YjM|yWd7I4g+yKVeA5_T}RZj?E z1RkxN(j&n7z-KvM9b57+{B>?AW7J zl86^ZH|n=cjjxZ~+Ygn>XH#p-3W{ zFqY@0vAoadYRC>)Bos!3clh}@t~;;^JsOp+G0~|!Y~n)2rOo~px$NT=$T&6o`$k3V zodUOkbnF_OOSGELZSh*Nc^Gl97=Frcl-*VW2yE8Xtk@cb=Zb8!xl?s`6;S{9u$TdH zRG1%T8rF~F6`+j50}>;4;YXVE5%$qV)++`p}fvPH+_ zReCHa0tk%xA&tjT_K0Xr>yqM(ppc>5 z(O`;(RjDowrdRa|Bqb)-um%5w(6d<`3l+6Qb5yABZMUGMT%aw>?LcIeESV2?q?@3blw(Rs`Xh33 zSlYkvgX~=m`qe(!6149xh)O}{7wJePuDgmmSY6b)9=>DUf+Yg4>1#grviw3`B_P86 z%=t@CG3bFd#=_izaPfc)ZW{)!J1(n1wnoy+fncF6S4`VG?}9KswAgx&GrZ5&H<)I2hS0|E-uA+; zSy^1A3`XS6pYQRZH9=9uhcEN&3xN4}!$RJV)SspO2)iS32gVy+TxDbYSj91``qK=C z(uZ+OZ8MwS@i6sQ!{^>7;j9uCcfyjw!hS5@1=jwRhFcmF=i*Ruu89wdUVyEnVD7F* zSTGvJxj}Bub7{0p5=~L({j^*V2HEphBaP7CS1zkH1e_g=229D9OYy19W(yErwyd99 zp4k+H8$WT4OX$k*=s<_Wclm|4gpDU|Jf4tcoxeFT5lYO{U9lj;PuRuZ^WkXm4n9i+ zr9f|kTq!0bu{-YRj-BNE{mRQX7d{YaY3RhP4@4!I^oYdI?aju&y3ig#_;lh1>Y-^p zu)bT<&qzjOSnr&psgrmRFEL-y5F_6pC*ww)Z}gaE9u)%CIu5T||e`B9yHVVZTD*`r+apl4npe(7;hw)(~pI25HiDK2y)N^71?;Xb8sxe z+O@rKCMC99rR^1(BO&#B&$=Ryw&)(EXr$p4RdZxGIYleCF4AQa)>ov-{or}))H8H9 z0G!_^>UNm7Jbv3vY}YIYMq2YT@4=-#k-L~mry&1oy0hbskn3e%XE>8Lb3#mY?rNwP zyNUO`iokYw;kBG%UeJEZ?G}IpRxLH)9}Z`On72dO{aNQ1 zJcqsJM>3z9bwwW|;a<>?xxzlaVwrM=5&+%zJB8|otSs212JQQaWl2^*;WuZ?d)fRO zirWBHhol|=lY8t%0_$OG(kBrfG3f2Ey7yawT1p(Q(BiEb9C>-GydhiwA^Y6G_AIsM z5sdT;1J7wXs+uCZ9aED`*9;7Sjo^5Uk&E1Zb@nRnQ;M>j9hq%m<%Tvwuo^dza^qPa)VbQXK>XzAxXOy}WiLkbjlI-#LWP_}P zk=@8Exu$f;a}>(6FyAd#=STzu+I0hJZsQPb zDF~&4_Ot+D4JIkLo8Yh^xO-CVMleErPV97eySlMS4wBVzM+>li+h|EbW&zpd=j@xe zc3+@f>MlpiEE+3O=_0FTE8>}v!r(>72^p7jE^~*AOW`w%Y{>`vJb;yY+80=<_mV50 z{M^TT&HK`;t*Ds67~va6+?;ig_Vw`-4Ca0jv8=3C^!_~elSlHq85bx+pw~O!5f_X8 zgY80`@KL4Gvw1?jyrhpJ3-7>r#`J)tOLW&0NdpL%QZKHsyUa|oJkrr^{k$3wf_I1< zSk{FSu?6ktMDh!1l0KMB3Ucx9iD=}8yFXsur~B;htg_Otz9A#g8(CMV>qxft4_C)K zAFMptMX;S9%>LeIhcEH%uD0N^rW<>h%fFvUK@WIW>+jFTvB3qU1%zE#%5{RNSUDpM z(;m`p3wFDJNU#ij$I)rFHt$jeE192}Hdq?X5q6wkMZYt2EWyL7N)lt-IpT1bfVL<# zCmi;uk~1O0yCSAs8k_;aPg8*zvU^DtV?6_Sd?U~r5-aFsuH5_YE&?Y4A7ME0!YwG9 zSysYK2hj(fyEku)Mdq-QZOr{lTqEfw5syO+NzZSZs?$(xYRj~k1aEYB?vKo6uUy^h zO5uUT-wlI)TEe~#b?3B@fE80ks^f)Ff)2>U)cnrWdfFh#53=?vZ+j83`&@ACh@|Dv zZ|n;hrAO22etEa6n$1)&(8J^M=B27vx>Bx&#dhos&yzi#K^S0x}4^CmY)rEe~mXaP}UQ+8vkh>8h{0*hY1sPCJc30#`(AfmJgg~xS? zn6p8brfdGRaIi76yZmbZgTftVC-%dIt$Ms41_!#WgP0NIS>!cDCfMidxjqVKiU925 z#tcB*Dn!v#Ilb2tXLh6%y(Cxb@1zzCBnpOrRzwhpS(J z2=fWZUx1VP470)l8l6qd7Yj8gI^Qz)@jt9@sxA(L*e=~3=KMak9xcXjd{W}n{E)1y ze)pg-8r|k%1ae*w0RQ7Q<~Exv^iOztp&PZ8_Irk0ry_Xd-22&F3FT-0+wb;n=21|5 z!G`bm^(sJA#Ff+_B$~I z97zQV!ad<9a=T`-uN+}}&+Ds^ghHM`HPBc%P+WI^B_B`vNyVcJ*0EI`D7bO-HzJ#S zgvED=6^(uD5I{2TbuC$GWv|l>a(eQbb-w8o&61*`7?HGnxz1%oO}~WF#LTI&$C~N3 zBo9@n#^Riq&tcXtwDGHlj&?^ju^Y5=>f)XD2e2_hv=Dh^bU1*6DwD5XF~8h zS&uHJl!#NbF%l=-W-zm}e<1D%JejiJ+u>L=VldP?=IayUhdiITei$hRLadvZaKB0m zf5<{1k^=Uqw|#Y)Q=>l-PHUy!ewY11ZIDSoi0--Z2eE zG}Gs~-w^M8svxbC4&^A#uXERUXj!CO$dk<6D3;@a&lvl6ov!$^v!x z2$!3DnzXZ!jE4+?Bl|tW4n2M^Fq5(E3D+2|Q6!Wh-B3q!fDahcED$Qkg^iGDSZ-MB zuH_xjpS_%GCo~e1!o!(mg^o1=S5&kT{X{g9%U&5w1HHZ(ARPeQ{Pk4A9;;1Wm_ zuuL2HL2i1DfNnBx<5O)nnjOLVA+o0ikv4cHTJ>Ij!eQRQq(~+7nhx5T`9TQcHp30Z zA;RIl9P=KEVfZexn+=@xCfiIP6|kF}lnmBsO`!8nyWg>$ggYCM(C_=9J>K3$bgSv( zv9!+5P%iq`M}-OT)Dq0&9wE)tQfvE36mQFA*xH{oq(}Y25yGG2=*8Bmem)K)J;sTV z`QAs!^N;a|0rASu{=naII4}y^dqYkMTD?4vw0=tq1qGFuo)Hv({5UWhq&h;S#SC}? zkaOu0!VdZ^O+Kk;aD--}t+r=#86B#=jb+VTVUXiJVi;f8wQ)lmDfO8@jP|114T&7% zixg@FYj>*{xP1(PM|rfZj%z=Kmt*v6!(BOOVsg-J+Qq$DCrA0=VKQ>Z;^4Kuozj$O z3qjF_0^y6YMvQ128~0{|<~Dz$$h&?rL8xe-C_W!skWTv|=p#`wW=dSz;(pvB_1E5M zAf@;RwfNAY$_hbKHf8<*wau`)2(>$U?QWZcDdh%*D8}eWvmO~RFK{cV^WK6+ySLpK zV(XWn>3i3xR>FcqQH;*{BiP@G7VW-k2v5^+`e33-)||qCED7F$?PfiC&fVn)zC+CZ z?Ebwx#jjZVP-MUNmsc4{EwrM`Ucls+=e7frI~oYdb%qkvr=d6%fF1vAi$fjJ9M>rD z$Je9(GIBL7aGZj@Xb6ngry_jSdE>G?Y)jo)dYY2o3*f(P*vG*615|x?CY|38D8#0Wz!f51IBgqLEHuP6)SVGAwk*N#%25-ch8-+2(I3y-iuVg7a zZr?{$9dEh3-~~`JsX??I{#FWRf+9PW(cZ)TEn;SfLzl$9O}n}?S%jok_J+ph)!1hm zo4>LiFaJtp;-qtUKXiuF-8l4`-5Mv2EiVG-x@MI3?04V&G&~}H-G3SS)mU*Ok-K9e zuD}1Xxn?K;eRLMjAV9%V1lFg2|J+yFUpIRD-p%K2FLdh2IEkVnm^GgnY&vNvu3c!7 zLhM)I^+w5ef6fDFWt$%Y#v+&y%t<$LU%;Q&3}IAwdVXHWFVO#Ju$^9~eW}WWY=Tf7 z5;G2iKhCE%rb|0I&M6I1#7ojhbtn>~{9e}5voSNHZBcKy z)!;q9%{xWz~1(BYEzXVa^ zj3Y0nt;?0_M;d-4__tuD?2fJp9kT(ZEu>YK9ww*lJ=rPldZ=${Plb4jGLe;3mr(a= z`uSwhR_fNpEg<2(F;0j_OBjo+kc1B7tZ=>@m|O8&-LNov5rh} zi5l+;o6-*6;ZqKxpzjw&_ zJKcq{8C?WR+|F)LL7#gbwpLYzbLD*Z1N1FLT7vB2s6LuM7SPrG zg9zT>&u0|8l4~&@;k9avCheHNyr?N3-K^}3n}4{4=f^Wt;Blk|$5}O6xgNG8shk)w z2M1TN;q$&?6!09FU=2DyhFg3>ESHm))(@u(?lU8a7H3_3Xietv{!TXo{F5L=J~V8j_lQi!S1dUDe+%aFa+as4MLqI&U;C~qpt_qHS8bCa!u?}&FS%#mgoGYp; zMGPJ!*5UE@UTRT)r^gFfeCi`pGz*lWMzk?j#qsZD{>T9xxbHj6BJP~(ITHTROijeK zG63*Ys+)*Mf&*IEzFfVZjjq4${dM2r^SlE1fe+ko$nD{D{|=PgM=&e}gy-=Ye9_FR zzPHNn0P>0ObznD)jS=_7RaWWw8Y1p&-DNEx32beF0#Be_)}{>c>E#xM%2Vwy|9nG_ ziZ~i}QUp> zLg`?4%W+hg@BV0}hifwX`Eb5br3DVa@AG`%&6D{_*6I$tlVPqtRkE0=_a{=)?JWOJFV-K}WY~!14z!W=NKS3iBx=DmGj$ z8fIUR;25nfXnFf#!_S7p2rJ9p==GQ@V%h~V<8sr!Qk@0bS?!uig_+c9#MnX5UNSj5 z^h&zjeWkUt{yFV;Y58hy>v>BqrRp?qVd5twr{XxjeIAVUlMYwGnGh>THxnNiwB07& zd_=Z4zVnPkG5h*#>so~brccPYPoeJ_ZKTvF+bI+f4tJM?%J7O8n9qIU-j6R~a{Cj` zG7!Pk&WlE%#%reN=vQ@anV=~YONghm(byO;-#;A;@|q(DC~HUSehm1ERS-i-OyNX6 zTj_bRnQ4>Yj7x@lo78Z0G-n3<@S87r&^nBdLWql0^d+793S1UMH|@C{vns>{0k9T* z@p!=dAm7*E#F^>2Q{SFiMq@Uw!pohh>Co05dP*<7cfd~g&G&m=szUh#R*%${CNK_2 zZ)lpA$Fi@+)>G*nqSJhIegpl;I)oFft_Q~~7rp!ovQ!bMYhL6gm`Y;Sai>m3E<7{l zgBU%`8g4)Rcf>27)BAGrDxYF0!zcP1hmYV62gDl&^;c?qoJ#Xa8dHX9Ag8E)}C=6mQ88d)93yfCVjjV8xbQGLZ5 zi1Gk;ecsg<{0{BK&$TZY&!EE44R{r(wyHKI46ulCLa28!hfOp#xPosEr>-lACse===9Z8_Lbqv{OQf9kYp-@jygOf4cy%7anFt`^dpC;h zMg90I-CoA5eA-|3qTcqu1yh*yqu?NnK z$rW`G$4O^Tx)0g>xwCRbrYiJhB8e+>db|8pe;y` zE6kDCmgg0;xmNZWd=4~Je=m{A)fECD>17qDyk-BXofP{D>;BEQJC%PM#}{Cr$RzrW z_K39$&ASf^G{uTR!d7B(g{VnS6z+>qeo9*Jnu`+c+oFcUe;0qC%|J$jc?4YG#3}(T zzp;Ya0solq%4yXg{j&q9{q1Ofq937HQVBqeMc;g~#VEb;xNokGCP#};tgybqAL zA$|_$P^P%e)K%Ae28>N`kyKjuvj~>QMH4_&CmMS+h2B;kk!~J||?73j2A~ zXBbJJ(&}5~IGD7Tf%75wt#LtjvWL_o+vP_Xo<&iNz<^Hmt{8+tmUX@UWjb+XS?mkp zbSgv%3pa0oP<3SU-%Wi@C`24hGbooIA=mj6~kmso#dpSXv^G{ZKK=^BQdQIEq z?&wIBE$83^hf_OwXjDq@fPP>N$bmd|eID;+k0WP)1Nvr0(gvKQ-*?r?ED%QpbhtU; z*MuZ%e2*gfpUBp!sP06=Uorjo>7CcGz?RNR`M**Hr-CsT0GZnC& zOU>9*Y^q%h&0?t%%5zMwM=&3&5BOL3CCnTH-o`VRG;pmBfLGcHH+0U*eA zlarF8sY_Wc6Y-VoNymh?+E%3>pa&Vz46(;I+W_5pr!}N28^0{!VJ73zW=1JTg-zD) zCw2*44T_`~ECtP8kH7IDUB5t{F)dZI18qmBSlGS$K%SCjulJfO_FL{?Wi+0H@%;kk zdAFdOYV8p@u#kweM_GooeCl-GE|S8!oTXXGkkW%53nr^^ugrfj*(zl?mgH9LP(2Dz zAY7aJS@=QkPq@c*AhhRgto=q{fQoiBM#E5vB^-bDP?)iTgVJ%w@a-WFZHw z8s(WZAPUa$^KjmX9De+9Q^NwZ<>n_Iz%4YP{=f3Q4-VHOL6NYRSuhQ8vh24LksCb1 z+ZQ>e%CrZ4az7ez;!o^p9D0WJc-y8|PFB{Cvhk4ei$hMjy$3d^cWs#_t!f)>Lns^~ z=KCj7nj8C|Ru1YVw{PYInc&P{NzhP=J)8E1j6d+aAy&?rV}Q+^IYv|+-t>+~XWZf6 zntPfa3Q9tb6{zhSf9@puc@6<#^H}XxhT(TJdMDZ;gOZQS%)ZFqfw=td=ycpOu}V(P zf>^yh$Yeg&sXI&h2Iwk4rp|C@u6D*8mN7gC6?qcO?`rp~=rv2F)0g{YB+d|zq-Rf`ULPXE^f;&=i6I^P8! zgI7UwBjS0uFpR=gE){&W10N-;=mDWMT=Hi?&p(U0qr>B1&0TG#i*5sIG9ag$Ln#wf z%JMwh!nni&p-9StK*B_Suc}i|ACjbbUz(8nmb&zp0X-r&#+Qo0fVt&z9GKT*Td5a| z7*)m4geUS;SmY6+p5;E$o<*}~>OHYOs=`!ziVf?xX;kOZoMuP+*S1_i_1p)cAWzkvAeW^Nnd%4awVK@XXVJK3X-L2>T_60LN@b-L=>}23DFt z9?#0vTv%rhw|J#!uDHTRnw8-&!ID<>R8C=oaO(dvSip6vB-ai{c zu&VjIlDY*~k`nLBrU~~o@!^EEaXb59-^Z2i!+axwRTUG=yZ-WyKA@WABiy&vQJd?1 zNgbYYn=6)lYn7ad?=M9GZZz~+KgW=JX7X=FOay*@fB@vG%)t z1^A@Xb@C`N32GXjjM+dgvul(<83i8nRGH0PiLrD_c1RFe)l2Dldt_;v&OfLgrC6vU z_~i(=A?%Mao*sI+ya#4|^xJCvf)MzSm4M&rp; z&U*8(o(As{zV~H)_;_6qQ=BVnoc@|kA|^4#sIOa~$|b|3(F3{Y(t(;1MBT6zuwQmM zumdO{GVYKp5WhF5d?qew%&kP0j*8s%?aEM}wdb`}Q(bRGJI(f3QM#-U(`oB#IK&}f z?v>hHH6L10X(iap&DqRd$o8{t+hKY+GpJM zDp2rTzp zo?pgabz!LydRx6T@cFSEvVn71bKTx7E4fn3eC)}7Ji@2MUzIMxsTxJ1wC=uUR(4P+ zN2HaM+de?zi&f(>`XVZM;Jp1L^fC)=E6Tv*Ki5jhhz~S5>*thjNA43`S~3AjPSO?A z1b;3qAaTG$CuGNXJ&Lq`1b1OP0Vo6)xSSoRK`=DR322ygOc|Y4Iv5#N0BKApAepht@KVEf z(AXl>qIhMeMypkfr1fg60Eovgkoe;$1`fbe4);j43H zU>Wd@(Eop&c{@|+OMmT%v1mdm$E6P!=peL_z|(FcQDw>(m{DBB|bOWx^z_RTMENw zl!YQvpGU7Skv5K@C&`HC4H2923!ZNi=R^`ubt+?D)DW{e{5+sXfDlu-CtXoj4q`81 zdyi%CglyeV1zJ?d?p#DyLy=F;WO@EQ!b#Bz+%{WqjzPd;zEuoPXOSo$;RTNEK+qCARgIbiV@jCwdHHUJrgv90O1k08-bL4ICPB z0o>xPyX8L%@Q+4HOUs?1nU@YifT+K90P7l=@u1m0K4~aJ?ZB=5vV4E+d7Nm4z>kvd zxYZ7bs@%6SE`gP9kucBAouvIfzSqhh1T!q;&U;5a@kt=jRG)~z8-UWqB2VJKvcpl0SI#2_c_@;iY?#8KNe8#RL_9XG~Z#v=R7# zm6N_t#nWF78Ts+EqR|5;^{*i*chnid7kJmzhzr}hpV+cdv^7KIpS)AJTO!&==?Q6> z(BQ7#e_{&541Q3oAPE(!=mVzWpU3e-zA5=YZ0f~p+EIog2wz&hfxJh2>mcm1>b{K% zz3z~d9)U4D?w=p37})<&ePgVJakDjlely%>#N(R7ucsUnxSSe3X!9rvYcK~2?e z`Qz}Zq{_CKb>cD>5m(YP^qGUV_HorXJlj04ts%;vP;&mhYrl);q6Xp$Zo7iJT4&Zs zfP`@!ANdB-*~f941E8L`YK9fkH}BN_=|{t0DypVKJo17Kh5KbF^rhDjVp0V5tRxOw zMG5xVh&}l#n1|p(Z#1i0PrM7s9pSy$Uus0;Bj{5~{;Kwx_;;;Uxt30Sw z9z-D!S1_3AV2!Ymg2yl(=6CYcLLn9h0`$z+>E7Rs*2*i``^0l5u;^zYHw zdp#N!UzSa1I1qLpj>N3e<9NX0bI~jZtx@x$Gt`NrlQ1YpqP!{(KSV`<%flkgM^!ZW z*oAf(M=VfTSlcaux)KAZRCcXx4>BcmfYgbbcRZ{QpXZ$p!Q*n&%<`TY$H8)RBn%{g zSLTsLz@(H!&xaoHEcJXnj}%28bpIr5gW=*htcx2Rf^g5nom9;#2&Gzd%W2~LrXt@x zYj+SX<|S)P@wWfIi*UFY*X*ocZ4R^)_9j!0?z>e=QE ztbXtO^C2jt?;rKLJvzj+u5#7c5@t%+;1RV>+N*i zi>Hr9*jPi6+ZgML@rStO77pyX-`2+-fbr(%PQyLSB04=UKbMK8>*|~vvSAYe#uA}H z0Gog5I3i^aHxpYMr=p>cGlMA1P`YvwVr_k;l;pW?*C2^aUKZOY?;H zr{JssJ#@~pbJZ9>(PsFm&eJ#~tY=$^@EPGLouapPKA36feDML~9!IytVL=e9AIWVz zHjga1dJj*xYvTcltO?Ryj;D?8=q06oFCH_03F_-$(EWr%LkL5`nLBy1FW;cN&K^6?Xe&fYUiu2#S{znb&)bWW6BX1C#rU3+4J_ z#KoH%vn27O+4g7a*HLb-`b1frxx!5;mMNm*68ZHY`y9;CdyBm2_He#>wyvAiOG92< z$ndXxlnGSxfe_K+?T>^s)`U5X(lf)+K^kWoMy;5%0v)vLJMim@AAo$mxu^XFB0ZHW zFB2v)oGVF-eR>7!-XEc+z-Z5#%bgysnpCpwW4S{6dyOj$fU6&@gJ54p%Nl<)D8Qz- zM@LAq-yIHXLMsQ>s&3UV-Hk0;(Y)H#peqi;{lQ-(L$6@Iw9@wVxaw>p=75Xg1-iHF&4>C&=2d+GQcpPQPt9&R%vkn6p{ zp2fw0yG|+NVj=a~E#=L(iMmSikkySD=gG-^-|>+Qa>3W84NT$M?})fp#B!S-mFEvW z1BKxJRJ8wl_tMvK5^z<7GEJ#3&oaUS}h} zKL03HPDf?n$zHcIa+hb5Vj<_yb_}SG%9$Sy=NqQe&L5%fe9#lmKk@%uxSg{{|AieCbO06Sy3R&!Dq)<2Y{&6N33#Wmh#>Tr} z@iO+dEm>C27f$v&m<%hzscOjQ7SO)-gW~S~wh_~Kik{SSB{SR?NCSSfqV99Gt&H;V z%n(wDn;7~jC~FHlIwqoye35uWfwI4K*6r4G2EsHE?aCt&`7s_yjK1Q_I}|8p$S}<_ z2Qa|8s}wvyO|R2mCdQVRra1O?4H7rSWN%fy(1nI0Z$-*SxF4a}3#t+KYW1bv=mq-1 z>7XZK`4y7GO5Zx^{5ZLWSQu`t*0i5|rx3wk!(Y<}&4;U28BwE5=k>W?NJt{u)@rsH z47QK%_0GM3)rOJ1bx?3i!u#%7gis4p`YpV01sZ4GCU=W_b$1%HCp-K4nxrFHi=lt1yoByZ#P3Z%8J|(6#NT}N+tYb zh%Ry;`AVqskU9FZ1Cf?PgWKYRU+7UX^*BNjX<9dkVL8Z|!Mbi;?-2PO#iy^YVEE2B z&sD0yxo1z1d*5MefaWIl0^MBE+T>f! zAXN?;6+Rtzp)NmuuS=v8@0R2oeQhyM_YnC!LzXYHPevYia0SFng3}bw7QRWu0Ptuy z)yI1r6r(S~#NI**TEDeOQv)2GMLiO}?6F5BkxT=!8%h}Gw*Y}N>7O+Cs)tn&bSmnBN>b1cLTQw4vzd%JdY`XGKuoA(%l7-})N>+|NCxy^ zX(tAy@oAMxPO&#r>HT#&q;Ou-B$P(;)B4`dR#?Y_a{=MP{D8JT&aMrN?R19?C}EnP zjb^#Y@rK9ZE@(gfFJS4f zGqQZaWsc1(ZQfTy16YwAvG(B8QH`>Qds#n*oY6L1fjw$45+V0>dOo%g4;Yzk!+>G4 z>_NHRj{`-`a@G1OyC}4+%p121Trw~ZXTOj8uHPZMMmSPnA?@DtUW%Ve zW+Ia#U7nHrj(@N7O23J%`@CY~I#En>+gnqQ$!=kv;`~ZL+9-%?FSIS}@K+sZd=E?# zhyV|C^Sp?#o|}F3ajh=^#1XL|Pu<|>O5{jZg1>riY#xEqy;YJs(JzVL+c=lJ-CwTR6s8PrR`St)iLKps*0UcQ3lRy1b6JRT!mTW|P4IU&# z0y1&VTo@!)))`W)wb=(=(_TnL+-W@9TQn48OiePM5WGv%3|lMp1ppY{n=Vy`c`uX!qfz@jd3tJ6<7CjN}KQbpe<2$ z(#LY$&<~$6JJ}t99Ai<-h;lN~Uh8KBQ^A@L_Byv=&O*G1I#jB|7F{kDc!ttlMue1n z9~0QX=ntYFq@saF?>ox%na{H(8mHkjKwJMj9Er0WD;XkuYM=)%VnxR6n(-}p1ziRW zN|7D+*4@JQq(wa!Ivg`oA3w(zqIKB@q@@IT4T3%32rGKmueKcaujX(^!x@gh{s7mQ zw(5FO1RM?CY1k+T$C*TvKarC8{Swi|Ccq0z1T(2=2!mb+%sk!{?N+kubkC>9D$c=} zo@s%Pwak3hLkUd|Y@BgNA0WIV^E7rp_IGx6_h11kdYl2d)^77n9o zgnW8gLR>twP?gWem+6_T#MB1_AS5Ayp(-O$dcH(c1gyA1Fk0xO?%DpTyzV?pn@}8i95i_!yt1lb9UUh_H`6PoNMZVXKmpX|B zVD+1BCb0cOaqbwdq@k;&SND$>D7U>?Y#Q*?Lmm@ATST3iT8GSg`W&miDc}D30cRSEtKJ%Tjo=-DC&Fsa%cGtYXi5AD zNmFlKw*b246*76<8N4OJqH0bCtiz_Pzs>lK2D9(Hzk`hzeDeOzg9{hF_XD*hUxBzi zhpcBG(a1P-;qmm0ShV9*09<5`)N<;30?h;(|9x>=*lKnuxT^T5Et;X4?EH1(-zJ3W zvpT1`@VrZA{Cf||;O3_3i?KK4e|e!ZnMcIoaR!rfSjHHj%yr74*?Yfh_0Cut(kna_ zimwq#$GC*0Hwq;RfyVgdsk4Wd9D^4@oFB*xNXZQ&PQ1#!K);{yRP+VcimA@U^WRl2 zL8Epmy`NQ}I|t-&fBG8<(#wk*Iv<|8JZea%4Q-l7-0plmd6#-W9BXT9B26{IzAk(* zvTaP-xR~?MlY4d?9XZ4yMZTeS17(tAacnBLrSMl18SC~Z#cRwJ=Y$)n$x+c4EYS4@ zPVDyMp}s4BZw&9=cQ%ix&W5fSOkft@kUA>Ve9KHj648&L5{f4$#x59?tv%~U0C9+f z%*ZL;_YOyT4v*-se#nk%itb2!iF?{qa=H=ndQE8?h ze2A=obC6upf(A^J_6<5Exx`6S4U6`kbLB*I_!v4=jnLZ+=PG+c!UT0qnBA#o3$ zLK`qeH*h{N5$HbKGlm_RKRQk&d18mzl?COzi|CJlBaV_)Th{$msV7%`soVXrCCL85 z0^!!|tK*Wu-H#oBd9c=CV-|JClqti$23+$RjAc3H^4Xj)!=i~%1wlvOe~CAqwbt{Q zkk19l9fvkRlu$im-d2LT?8{Nt!z~M9fn*AjqYY1y(xb(zCsNBu)jAGPItx5lU7hS} zh%o8;9`ockI_FuQc6?m%nz5&mdbclTK!keGF7xtDCD1ZgBNe3-OLk}+-QtXZS77LB4#BU)YLmA{>>(7wo0Vdmo^dcget3E56rWMU_BG=xL0p^ z#9h;?7bG(h@~GGSI)c|Ai6sLf;Uo;<2@q!91<+&A9C6)7f5we8Gq?b%!L@D0U937S zve}tsTyL>ajGFol7MeMAA6=%iNLl(~8V&8BDRSif3>qG-?yDRu#XpVMStuJ0I2AHk_XR-fv~I%c)SwBH{L z37sxC|2Iz4!bZe;)R-8iw?~!xP_;7$ea7uh%{GQxDXtZ_eR-kPZX6{Xodfz^lI*=w z3|8>B7J!NRbX68qN%XUvEKMwzD0u^Z2Hfm7cn!Gl1^ov@u#Kq@>MzKcR6Q`znpPNs zENZ2kmb!^N@J!XodFUV$O)|kePtfuAk?*Zob+Rl7ltL z#E2$rsz6d>RGYHl4dnYCMpX2zHxLLpQ+#xYFdLA?`3%S4-CGF%(qq`bU6Or=v2g1= zC%{zWl1I142xb%_IKe8!rnZvr->iQTO$_>*=tMBY(aZ4G1XCfQLAupnoZxO4x7HIg zGd$Yaz)GUFA4n1ji1keXfZ>Y-4tKqcN-)1u#Nc z?U^zUHhA)nz8D^NomEvMxw@B+4AaGoUY}l1w(Iix0Bs^D-|B=l(+)PAoqql;Kds0X zU~D;`O-V%?--9C++*zr0v03Tv z-FF@^27ZTOFFWi_t9t$X2|TXDfKUCTjnMy~GREWUhX_OsZSN-~XsCR@%qKgmpK38q z`&|;9Ztu(JU(J9jg8+fILpsqX3;`-TV3BV7iP66~eY8~!Dy6x@LXNaxeN&*1Nzkf5 zP0!_6VcJ;Y*KA(MKUV7}!B`91BFovair|B>6-a_S%VUDep6f2+q?@G(#+Ns2HybDn z;X_1+Ii+fiRq0mz z%v}nNa?*adCo_43@yFtx(YZb&yXShmKy-JGwt5?z3k_o)G+ntZNa>H~qz(2v~carEkifW*zF;avd1Wu3b`*`6k#AUs~ zLf4vS#5CYb&33I1o(5f_Yo0{bWIpu}=4Af8OUN&R6hbrN*VAsYBYj1(A}AoArMtHc zLJ|wr9?`zATzaBMJS|YOSr4)sb9PF_Y&~9N8`gx4IKz+wKeH=u zoedlHgp@2K335%p1jt!$VK`F^1zQ8>{Ez{jf67U92c@d!!N)Szf>QgddE80`jG*WVKY)t;A%lcl{$) zMnM!<$Xn*m>AcbW%7Jst2?groHTf7>4etM&;wnkcnq2q3{fb)ame2eLwux3bcgTsg zo)#K}>hac|@Qb~WFoC8uk?rk?24h#5 zo(Xe0ct9q5pKlf^KLv5js@1oL+s$Tf`ytIB?KfX!qYG}k12h@qU-$JdSJ5C&xif{J z#{HUkBLgXzg<9~X-*&n?F;Yiap_?B1`9}kYr9cYFaj?Ye7CMxyMxS2@V0X4gOPbOd z8)T>O4t;oE?x0r$LGk9NI&R9!brH9;rAT2f&5HnHXw1Q!ZA_2{Ty>=KFkM&fT@HXk z_Tvt^tm11v{sjH~?uDpA?qPfusZ259FX_lf1pp4dY#3#uf;!{Q+u26U#pKawOsAKh zrKA}#vJN7KwhwbX7%GKsHs#njwvQ*}aGV6`kv&9K|Dke&#cgB}YIG#OU+76bMQQGp z_z&t20CXSq4~(zZSpc@>yYhhJBbu|wO5v9DyR)LX4ormQmb=#8(W|fU;r@zueSpg$ zSFp@w$n^!Z!TMJ8d&HL+7UT`dy>o5)iy$$0)Jp#(O;HurP9!3LU`D8q?!4=G%kDQB zk$Q~`BI)6ZU<%xfQhRu!T|mT7<;FDDr!w=&jhYAymV2Z5djNq^yWWLk7oEnlhaAEu5xI#8$h>mJ_W zV+@vEEQdZ)V<_^jsYO-~1?Zs$ymHxjy=jxBi{vZ8OzAkmiSk&tIN-^{DLtm?Q=W(p zAs7oq7as(#IJBAi!0DkKGKt>hVgk6eWF`rT&DKbXBXA2q9#DKQY-;vU*|vx z!>y%|IIQh&ci>ELcvtXgC?^+@NkGiIUv)d*se0|StJ+%nS@Vrvi}iv>3&Yz9e4GIn z;o3d#`UQZbrgs&ijG%WZMB;$YzUnf#Yr5;9VYVM%_Dx~IsGC9rw)V&Ygm(~y6WM0z zA%zB5-a~#^nLJQi99SQ2L|h$V;6^v#ur@@W52$;uC|to^s$8v8;5`{+^e(#^b&2vK z&&p3xAclEKz&_y_&RDlL@66;JNR9hF2b&$9!61`)tUYdVuLf}zJ-QF4e>;;Y~wJiC%gXd!!0bS~dZ+Jmo4MTyhvbNq=ImH-58iej!$-dG*A9s{0+^ZTb zeYn3$L{Khp{9T=Qp9sl9M|cG=K-)nJiTUjG{l)m~3&3jK#iw3dAl6cdwUOTzC492< z`CzQ_TVK!?u|_IM=ZUCdwmcKD949SP_P_7JQWoUlm*Db!c7pLFg~ zH249Mj95j2!^KWohigseAH<7;UQlr^c)9ucY2N-9I|xTQZ!I4cHm(vMZ#3pffarT% zYLhU3A1(Aou-KS4Q{7#94Z0c*osw_t2ppX5-QI|hKu7rDgjP6c4H|`(iKm7+et(MI zeqr$1h8(bXR?}0}Yu<2=?MGJ-r7?A^NSKl{-m=a{z4i&huH`#deUE+Y`+mXr;9S2p zl2oKMQ1>XZch&im>c%C9yypmC$QySdfrm^NH1ZBlX`1gp=x`L7=F)- z@@?cN+ohP?ZR2CLs~>a7adwv&f^htAa>Bdgj<-IV%~5gmaYm}`uNR2pEfgI@jy|As z0>AAyKlkUdAr)GR@s~4&K;R#uQ$IjjG}Z?O7=JxoP801o;71&%&dn*bCO^8J$lr%W z(WvwD#0Y%F6bTTAx;91hkgy#rH9HlBR2kK)ov-HgR~u=Y07JVl)VDsVG?L-eI?TaN z1&bDTz$5PARaSwP>5x~l%W@FK>VH$f&d_#-oP*cpw@-e^1(za1=g)7EKPgyp_6vK4 ztrvd;Gx#n?#D7SG`x!^Qd{{NZjjVO7(L6nbcO>dXUN7MoRP0^Vg$Eajy?w-e2|~8Z zj9^y84nCfXgRnUV5r0S(AH&!xmHL+h93J~AB{$f8taIJ}Ui}o--TVn=m4-1Y1Q+P48sO?^7xKvx zpWyIT^yLjXmcV;z9CxQsv1W4S{%W#!t(s~v=i^V`HNSC?$*X3^t zbUIit3D*xaNEu@xF-a9ePGLQ%gF;K^f_oIGIdwv1nFc?W{Z!5Z_EV#8MXmE7gVHh_ zB(}KFutZiKCvFyxuap z=K?Cr2i(#jEmdt_cj5|ehoJ2R_qj&JHoxx-F^4y=W*PMZRQ^G=vqh7_s|%~a;Zh;Q z#gZHqUQ3Lnke;b?(y_l7v+Yv9KnjNuO|+4dG3CTeF>G@y%p!dZ0hU_j* z#s6``C+tpq%hk=O`EuOZ^u`ygP>1e(`u+?s7CCoDx}a`8t^u7Z7yW}_Z(H8q*GG7~ znamLjkc6mavb$eiS^x$m--}0X@VFJkV*)Wl9C|&@A!#wRjd!Vr*e%$d$D`hV@ND*V z;EX~~?#XAEKJl-&Ta2OXP2XdA?6kdtvxdx~QO{Rhv$?w@;Yg$V1 zTu3?pG+)XDqx1Zr?n>dF-79n<0g@*>8T}#H)^*&wQecSmg6wXNiZ_+h)G?{R0X}^} z*sM``pe&B~;_>#cH>3~NlJnBVhRDP7;j4Cs-vY<&&w_DV)b)`iZsD>rn%-z_18L4y z0n`oH+malPCb)0EsKhUz<|5qwXgmYZrgwS1-L>v8AlSyeQ)P7g{ns&)qRD0KB+35IZCJWe9Q_;LntDOWF$FaAMEsK z(8U7(fWSv zLp(irIh8t6^HutM&cpQr=J7?O>Z@i4T@RV|Z|Wd7y$586eO6hJKm{K$o((lz5*?sa?_^L8OiFRAbRksj0N|}x-L9k>W{m+43X)| z(2B=Y49po){8IdiS*AMnd(YvNjxe!N`13A3jhcy*pL;=RzJHObo}oP$vMbB=KTuQA zUOIe=n`xZ8h``<>zdZzzAGt6m0Z3#99Cd~zUDRQ_`0s815!W; zTL0RkOP^L3e!~ZTNzT(tRL^>ksNP-a?U)8$lvpP-(*}zo= zR0{nFnTs8g=P9@#g)xAPMZGti(d?VkqxVi%D({u+hERk~HhUhLDMUO$w=HH9uknCV zAkBGfyl;k!1Q*W<31gMc0$(-n4Tq8jp7T8+oMrnQ9l$JMrxP_-ANPa~j@DGfyd9m_ zaTP^9m>w6;e)$3q!Ty#68ETX0jYl0pS`nxINTl!W5S_?WoGc_l={^`-Borgd&%=YA zZnd$#5l8U5XpNbBzo%L1cdzh2k0PA|3Bi9X;!v9?q3rUF#2bx8mYi*HonpKOw-DNW z_#<*4*=mAyVv^mJh7ixAIx#oXj;M!jj*|Ayds1@l`DZ+d)I;|AlPx=dq{jO*J>Rws zVlaT;Ara8j`*=wYH=|oD193S(gcn?pYA_M5=h+9mxi!t|Uk5J*Ff;6t=Fd8Lhcgd+E0{uUi`4H{@P`zoK`-cwIJC|=8Lj;=+kAp zZRGVJw-J@@M>MCU)%CWvnytO-?CLhXCSFyXoZyvtHpl+-t*-`kXT1XR_HMO`^F}3g zkQ__AxJaHgZ-^0h`z?e26V`|nLGttqBbma{=I6&a%{Qj+dWrK?2*I__; zM?9Ji69C%C7v}}G_x8ShJ{t+tuyK)$?wi=q4&wzt zHD6%Ymu9NF1DsU2Bl5(OW`dai^2LB2Dd@-Ad7QBmF*11}P!GUM^}+9M#t0a#`pk0u zg#{;b!@JuEX$+yoUd9Qz?tbG{>I>&~5oom3&Y#aHh__W+5C!(~3@QcH;&&xb0Wid* zAXK($@QtNWnFMkl02Z?!uh;2^mTiupM-3B3q9tP5Xs}o*u^SYbQ3G)_%L{dSqlSGl zhrLhplE&LQBeTJ{i>9oC&g#fXCnzbh|nb!U3^R9W}8;%oa*W-SYUYJb%?AI6p zk86{w&of4Ul6XGd9CcF@tUFVqgCNDh28>U5U*G`AA9Ze&ChgFiLf*XdC^k%;)f&>$ zD`x`QumWNt3BX{!dWWb=oUz5mCV+MUGStRe1pcJ~@1(p@KK_C9)VoXAIM3yvYmL2n zd*;O)j>m8$S?3R2u__vl+5Y+V>L*zZewA)AfN6mir!EyM(yq3@a7CrcO4s+Lr)@NZdc{`g?Rol z`Az2AqpT*)-evc@%lS=^kPE&4Ji9H!=kXB<$i!*YZ!NU@F5+E|07((Wz8^`(zJKF! z36&Br2%vktc;rP3dp@c^?gt-Ci?gTc3tz00dil}2;qn(DFKVsZO+!8*zKJF)Ju!!jp-L7VuD#_{ zxx&5???b=hpHVnKIRX-BbbX(AV*3Hnv+uIbh;+hXJ6=pNGACctYk@^iCg^~pCMf@-s^){ zg!Y;iviv}^Pu7#k{l8#;_$`z zJE#LkpbBbJjpJO~w-#Gg%5U;Z!kL!WwlpECzQZ#q#LHey$L}w>rp-`1jZDLdpg_&x zoB{P1FVE*tGE(t4%$|o_1zK+B=S7wi4>C^UT0qfrPeW`kgX~1WcTerxw0xca?V`wB zIvdr)P;RCAi6O_7Ao}lz86$867hl%=AnHeBd|Y!RFAi1>|J30-%v|s&%wNHpzv-F= z@m}wGQ7(F1Rv;|8p!s%2uHa3sL6>=mjO+pQ#d(DMc|5<_5p#GLr{OJ-Dpt;yx3doS z;ZwbZtht=)ueskj$NN8)&a6pMEDFMZiB<8DT{e{+c|(*%Q1;DVKS}jObWHSAb*ZIr z?>*;aen}@a1<9Mc))HUS599RVuc+rB<)PxV(cWC=#Kp3W7(bLJj8^fjLjVz!{V;;( z=Q~AMgN2iJq<}Ekd27W6DS!klvdBuG;;f%8_JSxG3T>q$=8Z;RG^af37;GkrN8uEF zWvD}})MSkC)1mbvF-J@ODnLp90B2aVVP(V-S5oH;1nQ&@V*1$;pu}BuFIhV*`RQeU z3H{z=is$!vZKvD0tP`gD3UC^a#6zW#qSZpsS-c{qUoyoRKVYhZ#lllo=$joA-VPff;{#|$vxiM?p_KT?8&@|{&SF>jpBe)!m-`o7q(WxBE|zO=A`ShZ!dX7(iIqS7bU$2M5itw?T|YXoXGGKK+}$vcKlX*b+3N+<1a5lXZk%4 z4AEx(HeWTwjmU#$2sAtaKkIUX1Nv8*iag$&Hq`AVcpl{B&v!M~2UOU2J#F40WWn>( zy*x1v~F0{E!0S1x4^{Jn!(mU$PvD4tcz_-r#=2I!Z_*U>~@p3`l}&;PjOFHMb?; zhxH@2Pab0KC;GM^1JbEhr;nSJZS~KXuh8@#J_y_T!StK`Eh~tithT!I9r&(HkTafS z4L5r3&(1iS49xAC>Z6sXOVw?gH#l$ZX#t1ZM^-%O`k;(FPUL>a941w@tU-C8?|pl9 z$c(fX%y!&kU2Y}DZzmUs%D%MNKr;gPgaU4KT#Y(#8rm zkj1=p4oQu-Y__grbp1eqTOh+PpFjF3-xu+|!JNaXV)>GwyMO6KY=2?>JnLOpqzc%A z`}y_&#UDF=aWJVqFtg$--N=DX4@Uz_)a7wN`r9L`0VQRo^`$LZ+KR^}^cx+W#LJw~ z-K#FARHo7hLtX35)Pk&Ja{MuQ#?cKi4oF9)4Eq>p(-Bxf)=F^vvee*Li#HA-F`=&< zA}N%<*TVXAOtn{GDP&QFr7H;0_)B~Tonl429%@L+PFjRK9jo?e_s#t9K54PW%M#lTPT*0mA z=9R^mW&P^K72X$%!7dD(Reg=7VtfqyxC^cMaJpQini29hD8jZ!ch_o;Kxngqx9%w2 zHUKv(Nc;((e$b=M6VChPkWkpex5*fI-JYUS_XqhBj?a0{j`y)G-vhUu_u`1^XY`}3 zWDKXiJxOpSrwB!tx<9Q;Vf>H{hqT7Gl-VIP7@47?XOMmFdlyOVBB`imyh;%302jnx z#Qbo^D$eYs079~DOAhj7$E)6+4=)2@?eb}yLQ;&Lw|boD`yMO1vOm{!KW?5}SQ~d= zn)1r!LyG-C!vK+7J@`73!`Azm6%N10>G?b+{0%hLXE9gxdnL z-03;^km4Yi+UPD3CW69qpK=Jy8rM75`mFI(if)-c5-;%Ka;uG9Yl~hj;((5)%=Ab5 z=1ByEh$lvDOmQyon`Uhu752gOt0X*M^xIFq@<1cuJA0BQW*CxS)F|6JmX7JHxxtma zyz8eHgMf<;SChG)l_2y_O`5%o46LoH(OCSc<6f>}mqSUB-kqyij_O-^`_s=g&=P3O zmTy)~o9*mEDhDJ?W;y58jI1G|47jxTiN?lo+b3Xf=U9`DlD`l^4-w4{63x4Q9YOoM zXGaV@I-EmMmAzYJae%N5!=?2g8z|Gm`k0L#EW!0aZJdfmaRDAIGdxDVO$bb(b!`GW z17-?(+|N4ZUvz1P3{nXVJCmEw?co9emT5*JY#!)V%eNza3boG34<$IeE*S}sV_!P@ zGWWkLB+K6`+8Ky&z|No;pzu(?OD1OmQys?h0n*h+3x<0|T-o$>1csITywgmkMO|Iw5%F9G?Or`X)@E%zOi{SwD$ zL{J;d1rVNV3v_K6i|ovwRDHTJSGEIoWfCs@45DxlsTf&`7NBqxEKb>7ZQ+ z+lcp@h`;pQJwkGvSbmB3&p3lshsgYMHf?szIJdDMMBLy z1IG>N%mCL(!kk8(Hc#IcA7g6%M{nhEj=^x+-4GL#I^ZO_cZ)|t5#%XC&W=&$PIycv z1MP@~_6%x5Jf*e8nGgcsTqhU!vHY6FyjDbnJpD7n$R`c_%cOrwm7XFB0M2b)$?Ch< zE3im_f?q?sh{ny83lyTgCJh0O!Qp9z8iGSRz8GLMe|@P5Wfc!pwDlP@mpw6RGUS(v z9H^6jR^G&fk_CajUfE-jWB1h)3@Y#U$>CmwB5WIju&-AeqYe za5yT3cWAva9#DQ!DT3g#kT3z>lhd_HZ{(r}7K%>PHvvy4<0-spN%lCC=XEY3d|qa; zvA{zStN+<5((u~wNVg=H{k0Mpr<&xSlN(Uc9IhWc41_lHl+*0^zR#8qS@N*;>*NuK z+eJ_ArD~P(_bLX0DO2X9F|ZC>uNpc6(cQsD(O?&0x%2fZf`Ob6Z{Hjd!>mLCJi^mp?2^3WL&;uR1p4z20*3lWE0z=T~n zJbzA)k`F%R=e^(8&4AWg;AnO$28N0JXM9aN3=*5Z_suU(3{2O{SOH$tKb*3&Gd`2* zu9S+?ly7eWIh6B`Jc$mdm%KiYcqu`WNdqw-stGuEapek@xN6~XO^2ud^-s?)O36Zb=UasbWyU5w?p zOwX#2Jn^eL22l+N%|ZI4A3SOH#=ap^lMMkqx2Tbs*EUq=0BTy)TAVppO&V zpwP_(3r)L}VUxi>MsEe($;O(`SG$HNKsMZ$Cj+&J+3~XApyb&!G7`a74c7D>QO6?%B_1vXlu{&L#fK}HBhTADiTERu^VGfP8;Sr&#cJv`3z%rHryLFvG|mk|9l<>vlydxYQ96}w+*LKaJ*`IVsv zQp|w`WU#dn;@0^u-2VbX8}>`{M(x*c?6y@GDAfSUtPC+dBW))MvOx5K;BM>14rl`j zXZZc0bPjnmxsv0MI^L~ooj?)&yco==-=1bwA7SgN;as^Wy>>JW2{^xC{-rgAAEjtC zBM`e5rx)#L6?gk7=T8+0urEkeAW#ki(*;LnetlY~6Huyx9fJ<7SrgNitLL0_m~d_w zF?O1Z`w)levz4sgG34GsoadHYP*!1Ic6qY1+>_TH$z?E*ufV& z3GEVM0$a*pLi(M~zw{9|EN?t(EOt)XB9_c=2zCnlsO{tK4SX*9LIA2@+V-^7gHg^SD&hmsBqbE;;Nz#HO<|5gm_&!5TnDmQ7fG)t%!82511d#)!1_h3nJ7<$ z5SxWoSNWxKA>9HJLW-ip0})^V-s|RizDJmn#%Qdr!FHj|GjvT*WNMXp+~hHckS@^l zm@r9of5jM?{l;JAkw8;Gf>tZ)vF^Qo?Mp|-Ro?&rgy|5x+9O^20zX3VQYD9*Qz9j9 zIPiK@(ZG2w=@tP0E-S#yA?u4-It)mI4QcRxUCK4KKuVPCC& z-Z#BBCI{O{!R^ctv9_AxbF;qdEjSHc8S`S%7&27Rdl{)}4N+fHbK5u?=Pb>Avu}V@ zCCFi~@MZyZsK$PCnlNQZ8(V~WqlY#O9N01mkJSR)=Jy>jA8JhuM*b-B&ko4_3J|AqZ7++L0B&w;FVdWjb7tN990)*?>GRA3L!rv%E0 zljT&H8u>2`q9@u9PX|!ApYZK%`tNLZUCs?YZ&UJlBZN1Qc4j~VOsTF|e?p4cK0s+% zl>aq5S31#NirYF!A@i*X58LruQWL-()El}rRwR|Rig2-g@m@h7(Ql={9|Mu=;%<4K zA>|ew^0rA|^T&Uo5rrK1hln5%G?Ro6wXczIo49+hl>+Udrl*l2AXtIRk7VZTtA!pv zI^X~Mci)qsdJC7OG16~7SV(;nB>a^)->8#uk4_2SBVSsw(yg57cDU4fqSUMMOC&%Jr-QswgbCc zP9(n$Z1fSJ$9NUEY?a_0;htJ`s8pQ56W|x9FGO)T?-z}To)~_+AN1}uyJL@(6I>h7 z1;|w8HJj`Dx$X}v(fCM(GfF3O1%)CTwL5~FI}*6~&*f!=lIazicfBTOWL*9tyb$o* zLU4uBA(>w*rtQ#0Z?9d{2cLYgETY}*kNcuECEB00?!lIEYQ~4T(Y)wJE>E(KJNB_p zARkuRv8z13_a)*yl=#mL>*MNLw1>P%8TcMfR|cQ-=-VH$W!p)qM&GD2GVrjjwD1Y z9u%GT(klRRNS4R63J1y-9gdkeYjgc`iwGyU3fvNFZ%*Qh2|gt=>&;icYDIblmIc8P zIB^5{Mz-mnO`>k@wU#h2_;0{M*!-)xjOah2tBDYI8szS)IO*q{t@8K_`zg#c{ulwY zd3O-Qd*f^*K0tfs$>ig4_s8SD8qx|vCW^OtzP8=x+=?nlFUAEK=_H4(tz|?Q0G3B? zD9eldfC;t`+4{F8UBfpo$j^RnOWXqKY@0aZt zUY_UY&k5b4pw>&v`>c^NoD+duJe$lbg)t7jK& zs_h_2+FkrnF8}JsJA$F`cOuck5HFG-`n|!h@OcsM7cy>$70vC(cFHAS2P?*Y3}q7? zJB*?O4&QZB_0W9I*wVYb&;{cxO+Pp9r-%%wqG#3LWV`;0p!??Pe1C;+N}E6=$r{zD z>(&=XEgdMIMu<`X#D38f#mANB`ya1z^E_d{&iZz}11<+1)g2=HTeHucSbEj^!diyH z)9-=)`b_f|>q+W!-=;+bs!%B!vL{>D2>VU%Q8o!Cr^m4W*sUjlmITT3MWrYtn2X!%!CM~9P73=f+px@KwBiG2TE84+E&M_tr z8#^MMYr0u6%;v}|n%#HJp)a9WIgWIq*KR#a9A>mCt6Z^Z4HzF-eJ1Fy@qPoD$01D| zdcg8T7(RyYvxmTSb)I2xU|}@vZZ8T+w^uRJpX8PI9>=|;l-3?B6WLkh>I}vzzVh2k zYv+2gJQca|rU;(}Be?ztz;@wy)()T5$p7fd96F`@o7E%98yPghC6D&A=6OGJ;`1y7 z|1T>9U*cslOTKkUp;HqdFtdXn#@8a0u=ia39m{;*Ru6mShJfP&6h@1!Bd9{{(Q65O z>vmI@x8RFcSv2?Y+QzMS1-h3hQ=4^G^1uDk46t`2Sx5X?(` zy5aVtCG+h!S0X>@To7uDnp!fT+5MGYxbq1!q+1Ck_g&F+ai%1nx)CP|Jr2ipM+sHO z7AhzcguP&>Z?WZyh!CYJhkl^(y8X!x_HSQs#~-g*vE7^lxO;7))`r-yU`ZU2pb_mm z+ca8X;8pHT5P8N&Kdqmyw``g$G_B)MMk$6-ya6dAh(*_^&u8%sK}`&A{cI~!fmYQ`9QHH8jvn-ZIj;>7 z??bb78}3zw5?+!jc$%5?a4Wq1NQsH$Ca9>Af*ScF#yq;k@VJ8HNy$G{7l37Yv(zJv zgfAuGEjT_94Ey8dCmZ|yxLnHX zL_lgPD`xCO6*c`{s+xi-Fqs?j58k!WwfBjhk4NUWf?sEBh9+eP^7J6h%P!Hiy%`Zb zjj6e&f?9y(4mCZ3Wi)@{YXLQQtx zgV{k5kGD$LgkyS)`b@Z$=zb#sz z(PFHk0{b~z()v8`1aZ~R0~s~Fv!Qv5O<}e5#vADGfkErnF@d{9Q&{o=MKbf`X~;D5 zuD+HSnHC0Aw3kUNY^1=n_AnK4{Ym^23_2h#Z7uQoI6Zp4bv# z95|<%L?qwh5g!jk2hL(~BJW`x!e26jl|aY8*N5#72!CJ_04e<(f!oNpmLH$dL=+r; zTVAog5G$CLR@FabFz>ekG97w>R)mLG!x@4vbl2hmLL~mVr-2VU=+DxB`)Es?BqRm< z7#!%LjGe#8`N!OVYxp2{hk|e!tyE&1i)2bRgx=Rqq_6ZRJ1l0I2^HgJ1hE2!Q_>rc(Gv zz|rw;Cm0UzqQYvh>igq!F;$3>pQOmn;obpRrE+DF%N~1Nf-1(1QwsA=i#Ar+6Z+zC}^UMaNbvhEmY~2PEA0)3Gda&{j$ZI*+5mY1>RA1mx3tP z-j`HRzZlPKGvzRSJ6o`dQPSmmj;KIA$n+*>95!9SSM{Iu!ldl&HRrw!K`=zjH?G9- zgmn9Xxwm(L1OT3fzkz+EbNf^u)p5|4W)tQ7VnYs*wQY(BDNL#mES|ofCW=d8HV6M8t(F;EOI6;C|&hGT}~tF~YKLso5U__1UFy~HS10oea$SS+lDbe=s7cgORLRZ(|PcF85l|7C7dkAjQ3dMw< zh;T&`w-2v9E9@l83wnmfQ?j%}x;Hvv*eNW3d|8;I-{*|8M>3G4cU;~63Sq)dPKjl@{2AmrMhQ>IG~ zy{kwbfN>A3&E;{^`Ms=0*+d@OH%wCcX9E zxRr#5djgrSWV*6;0I58O(+C#V>u`Tc#j0%!2q`70Y$QcruCs|xAdLwB&0V`1VY^=_ zHQ?X1QTXo1yBMzCF*2$Cy8(X|Y}bO>UsOu3rF4~2{sZ#Gud!v9(L|=bJ+DV$W7{5; zIT__G6D_C?RJRV`+qDS)du?Mr-dHvqnd%Fs^qQ5`qo6`nyFu;X@-X+*7gZ8&;a9t* zD5yWCT+UQkRZl|q-QR5hijx5+V*qmiFC=NN%}Jt*FLKYq-@OOfdQ(*AvdHV*puh}r z|H11c9)ajRf+p5WdMUf(#nU69475x6BhDB6o)frh+6r**IR@tymN0qZkFcguq!f1p zQl*i7)f+Y(eDFfuF8It}&l7~7t&<=W2ofP#1pRxPpw1OAvRlYKT|u##*+wY=0G3}D zxRS9CkLCB`_?{ttka}f}vI>LrND(TJJ`9l+K4`kD)ZR8MUJBJW4`oCs^(jSa?dR~O zpn7Q|jTAp&)j{|P9OxyW&`D=d*3jB@J0~~^=2t_cdlv?xUxUi;Su?o+*ECwuy!e?< z*;mnu4zQ9NZL1bx4(|w|9a~24ksVDztrhgHiniX{DCj2Nh`jSWJTE z@%5HFjh<6jCs-*3`Z-)*dL4Gi?$!XEYQ2ZZ+OX|Dg%;xv@k}*|F6l|%YDiZxJ_z?{ zk~zEz$agZ;R!MTPE7dX$b?l<}z8V26*y#gM6MVI~lzmycT}U}gsT^y{Z-7r8d~ z$K9_01RCknSpn6_4HQ~K_Tr_goJ0`w@kD5E`e zT+t$We>|UTe`Z0MYIwQYd!SOKu49O^6n^GTHP=szT6lys3aG@bF!?k1!{TQp8w<9L zukv1hwXvOmmc_!2eQF{BAuislx516`9;U~+bl-a+u^uu*%ejC7W~8k6_HhNV%f{@H zVo&J{8xa4;Oo3ppXtlbn{>V`sSORD1`27Lvv0SZ=33KBYBlSz{=;oQpdrwb4d_`_Q zv1skL`~n-R!&q@Q5nEsr-$@SKSCEMxaw?!t+xs*~aQ@}r zS#=qchmXW{n8(k|duIfreA$3Bzsz`s>v$RBY2O$yHa9;=SHh9j*N>Ck_P_h_@wx!^ z7&i(grvPU2x1oMbAPghJ$Ma zeY3XDbu#pC{gzI~Kmz~AnlGTWX7Q@w#a24b3w+`JK>8>+BNBj=y}HhR2PrL4_{2!c z*Mo7|n+VQHdX9kdVJPkc{4MfZ&Vu0y^!+L793+~9BetZatDODI3$d5VuR+4+`*PcK zdOvgw3HL&*OThSVeOY`s(Tty6fqq8{uN8A{d$4=uc-4~qMy>?Cr)?=%B1(h4G^H1- z@zn5WGe*tz(FQNd#B>24?&PVeVkNr z7!&L~e;hEhyrI$m!~cum8)TrasFdhCR)4k{qy+MTIw6Yi>pg?-0ww!L|0wiV{J~rq zlEmRNQxDZbn&+M<3$=v)I`u^-xifLu2;4n41^1k04qzbTNwYS4nJa6@{;U^O_zr)T zBl0qGxLcR6t}BbATV@`cBeQo>(-1?~<@fqpG<&2?G626@rx0J|M@@JEi00zYOmKgT zZS>`{E4q0$%tz8a$LT1{Z^-p%Ccn|_;Zqpj_>=*xNP4~rT`G4nDV=J2COzJzB`9!i z=;DAVFjMomX9x$E+-#V6SJ{{b_p?8(?y@*8Sm9QD23WZS8gtvhQ`qraihMZWHC_zTkoT$|i~v zVb5JPUSIp*&rfczNi{dQA8q|t6cCj}$FS=y$63Q(EhJ)1j(#c773ADJt030v2{IIZ z27m6WTa&*5n+=>>kr6nyEBSAD%6K<9x#&Hs&NNeo9Ob|xh-`S11&@QF;Y<&&_ZR78 z3u=!AE(K#?Q1J1>6ivBFrjeHKJ6^|Bh>UPQMpYvXsnJX*(2vT%vuW-`U>FQC``TiFj^)bGfz+n(+&i8>dD+D>3I6lTq6`YOxy8zh@ z{@x{hZAq@b%fr_tkR&~Spg`O=$a8qIM241s3U4dz?dT8Gd-Uq?8#U&;L}vCYxkRs@ z-7{EenqxKX_2y}>054}3;In7dWJcRr$?=sTr~89h$ro%)-`<(ywcXz-NJwO|JN=Mx z12N-fV0@*Pyg(Lw%3q;L2z!9B3%Xxn@p(C!{XlRK#8{ylZGYGSqt4DNnlv54bz?gb z#`1*3$TVUZbV+ef8OfE=kPH(1ZMdcb-Vag=H=Nz^N@<^;t_J}hrx2MFT;AYjo<;AV z)h6aUpfGsS^TShp{NZx=m__}Ie+gc_Ph_Sx z-y;`e5mI!zUWhPAFEA}7SW*Pn7!rE*=RJ4?p$N{l*KZYFYTY4YMLWVfzrU8>A^Dte zA>y03wTn`9q!2w|*FmvxdIKV_s|v^#j`lb;-%A&A$pBc_Fs`s9pA*L&>P?2~i7PXl zZV#ktP0<=(9ypt_=e-qh2i7rh)B@q5NN7%4zyF*M^vQmWh$yxEHc(u=E&THu)9n*n ze25k=yGp>?Up9Xk1?&DE!2vP{71>|=+qYj5@r#xij%Sc`)~cW6I6dCkdU+g~{ex3( z1hVHhD!*fU|%0BtCmnve+YvANHL< zR0zN}zRU4Lo7^R^ZW}XZ5RBfo!8nCE#&p;HEe#hqHfuJK7HWS(5#*gQ)#8BY0)<|G zIX*dSVfE$hAJL9bUPV;dB^!zJ0l`t;a$}IaBsFiq{`3$0l?6P$Vtxvd<8=fa_*d1O zx!A2G`foTY`R_%A(qvSz0$)$j{`5!}Ylg1@y&Z61`okhzS+Em+#!j!CHD)cD0lj)F zMEpBQ$_SHh_qPatLR$Ua4)xx22tMk=zqJj51PkP!$RHq^YuP$b&TReCX3@ryP=evt zQ()Lp_G<+{)#o_DX>ic079Rk)014~bw$tf|{-A^so=ly0i{pSm9pygPp(zI<6hT$t z8@<&byJ1njl75X5_X0_-Y+)z@OLOGFtCoJ>r#)u!2h>Q_o*Bf_Y?4x9WgsN!QkBDS zB)MaC{7mM4$YGdwVIOa0DOKog-34!Jqt=56>9ES_0;CUGy;tf1Jksmd)a1`V5T^=O8tj6-uVDeXiMC*mq7;+a-IQf)) zzO^MERp*SNfe~3YmIyridaH59#T$5}=yZg~G!r8vP2oGgXKV2hy3QgTK^Sg4Lh4hB zbr0EUjRHgtwbPqq5k(yu3crwwM=R=ncsVlf^wVjza&vshD#XBs5}q$?xu7@w{y7;T zLf44(QEQy2sM{~qEjvsW=#XA5?A3RS__d>Bwg_gxtH?arAdrg@Hv6I5V(3AX9u@f% zNaVIC1Sw-=78u#ClPp|vNwnzo4iX%5u z*nIuAM6Zq=9~-6egb{yYL#xnEWnvF`M~4e;0YRmM_nq4>WK-V{#F!@f%NNkL5aQ4- zNFy_7mfP3o?`R+z?sK-c2wV)ej1z1MP(o2&*+E9qM}j@nw`Z*T>=vkJGYrl_O(x zn+ooWL%a)8YWmT~nntb{%^tP4t>K2R^YdRmX)sHbweKBDnHtYR6e>ruMZ!kEy z?(0u-u5iei(>qDBWQAb2W6(MCBcUVSv4}F-Il2WH?XD}2;roS&MC-OkS>Lv$FZTW{ zey@i0=bG8q+Q6t!HRW#m`7IeFpg9pr-L#V8PJ3PPmsu%I33|;X;;dRwQ-`?ecoX8K zzhII5m*@R+snrZbQ=q;cb~~vzk=LjUiB(t)bl%Y9s7VaU4*7Ss_AcmwD;;3i-|@bx zwGxB>dJOmV&x4MWua@es$F&XbXQLwYNj(iFayTMv8E{Gr0qYM8NH_ZRjWnr8!YAYV zYC!Bafh~n-ca!u+>@mVzw_$X$(>F*B{zb64mEoxD#f4xtV+j);0*Pr#rf3L^yQyJ2 zW$)kq+kD>i{S2HckV@&jSro?C-e)K?T`6 zyY1fG28#k=F5G*d3h9nc%<8ep3G!+pF+eLv%1G|f_K$v!2Q0t|bJZIe=W8_hpwb12 z+Zc^1EicIUKxFmb1)Pt5UKGy_AIW9&#tt`q`q7Yu;GdX0CrtP@a&kUQE%#Ry|3fPl zF$Pc$OxI}1X)40qKTsw*NR*!uO2wxhE<^VXqoh=)=stV&Udg6vLx`9c#18}AMx-UZ zBIC2N5aoM#?gehuPu0QngzKa(bfa(kj(lc}BD@{uu+k^b%4#Tul>M=?`}ysqSa>ir7H)30mvcs;1#48gE%lJj zR@D^^cBE5y!BeiI{tEHWA>2Dlk;Z9tQvVN(Wh1H%>Bx{I5M(_vn3}3f< zdn|7P1e5`HIU2HQ*zoVs4UIzPI5I5vSpEI~*RT*NCV^nPO?75f4+51zpC!H23V z?{(|e_jaeq+t?5S!-^K5cfD@NETY(w5fHUU05-an!3u2{=@h$rhq^h>50|;0)$4n? z>I)`}l|Z~L``GdGV+EX(bU)b}0lNVI!_)g1-uuNH-Ego7s!T2@%BryfxD6>sO#Qjn z_b?g`Cz9@`<4ZWaRFs|Nun+oc_;6gCG|LZ&pUQ+aQIJ@$F|bEiR4KO!0xha*lgpI< zH09)*7F_T$GEC~G^L9I6Na=|t+H8h-!GDm(^c?q1ir4QG!jak_8B`xIwn24XhVd+u z(T1e{@7RU|b{{mf{2~EFd%_WS9)I*n?JO{ZkaywuSMaQqyj}*iK9eVMjppDxxgcX%O<2otgsznr6L5cEy;sbQwBMed|VUpXw zbM_Q26&OYO@u{98tn0@HAvHhx^!B9IqUM}+JnTjPQ7vzIyl-5c^L?GY=mBZ$R>Rb> zX5YwL5euTmFZE*#pmh}&v^U@WzP$X>GE)y9n~C{<_YWHzPc=P6Ku6s?VmTp0Oz>sR zx@<2lkbx8MvFQs_bG3S|Ll(&e5_&zPhfY-oMMD^(bfV{*+p?$q4xJ(ZUU>tJn(JLy zSQt-sxwYVxylW^OUhM+QkKJ2`)jr$fuA_K*d#i~vO9csFz2GZWlt|7 z8mh3iLa1Wcv=(@Es&5EHlmwGEUHNoa4}3ah(SmdzD z1wzcEz$mTSfNL?Xfg7BZiv8FrTYJWtF5bHp|HsNSGQ%$iEUwBNB6(v~z{`s@+isv} z3Oz4jb%Nv{xHqDG_ih6)js*J-h=M6JO{KDrz9BzIr@Wf;s9tVWcdF$jup1W&PiurH z)eyU~%8@>9e+d-|Rh)n(1ZkAO;2_WX>UY4WB|t&q(R-D0UrXMAD|N%Z#@B98b0iLh z*wi6}$A6GyGW0mA%c7wD#IKQYvFP_^ZTc~d+1~vv4Z#%jl=+Pu?9w|^qwxb{72h^M zbq_Y&pXN*mi;FnAK9qGN3SQLHe(uDdI{I0v_nm(7OA*U`)Za}%Xrq5Q94i)AYEx=B z6pZ}^{Z&6WESyiz!;xL1$EyfyBAKF(?;h@nW?5#rhmkbtr|rXi0#HF&4k`!xy6ABG z-cwz6P@<$2e)0X6LoZvDt~lb~H+aQ0UlgS}9FtGCG{%&`AtRJ4VhdFl^%MTN9G91~ z{=|j3YS%u=320KWDJa*A>?KdYd~RmB0ti2~aX{OjT-UETp#x_A(7d zfYx7a_FLJofa7BYCuV~M9enBKbYRFCO894Gf1)>f@~lU@&>L|^JsxOU4M8l0z0Y9a zMO+EnUvScFDj?Obd!syS`-FY)xgF#+;%pxRC4tWM${#icdWcR7&(g0*cu*4(GI_Vm zgX7$`KwP=eYz1;~b~v4I^IjiimV`$$#_X*v3YiwGyQYbFRLsMSp1btY_80ohb4L{| zt9s49)&!q2e)owSd}seaz%+tCWapdTM<&@1C+0nW|277cVzF~o@vCGV7%-`FK6^vw zQ=|tLgx5gYgv59Ew_hmF?B<@3G9W=G76ZoRByaLgou1(nE$KNe_IwmG{35^JLiWva z5eA^rG~y+Fn+ArSCeYz5+iwC$+7^?Hm5b(Ufzkm3D=V-c@82PSwf3+TLlvBU#w;5L zl^HleFyvv33IE*Sr3;Rpk&w1Ct-j-JAK$tdaH44Gny~p>iu`vd(zx2`sbv&QX zGDO+}F-ABKKi|*6Yd%_aabg>yVPz(qKuQ{ncj|=+s4-UgE+%RILpCAeoQ&Ne-f#z41|{ zm_{MSajW(9h0XGUX$BC?x(i3Dv6QYboie=H!t^oHojk2TSyAG00eKBLF04;gmHvJ_ zl}r@oJVh3Z;yOF&<^EYr@ z5a}MH(+}J`HYx_kKT4mk18`S>ZYj=D=amy$z}>M{BD}Fl5%nJ3h>Lxp0W5rFvg_D| zkGNy8;3`*bDsrc2Lgm`T=Qs4quz^Ruddz3$>6{P`ZVe`LqO6s@ngb=f4_QTH@cR4% zF5dyd!ZsivnDO41Lt-3yJw}Xr=7RRF>|Olvf!}BRiC56n4xU&uwgnxL=sw0UmnM~D z0hqhh1X@O>g#hy+0R~^V(8Rv8*JutXp2zptJ)`@6nQ8AAm3*R4ZJv|oq&OlHcy16y zuqTkE3+6)_&;SDiHTpsr!=1CF7lP*D9F;eqsrFM3{wxTfN|gXnfUfn8me;I7f0Bmw z_nwZQ@a~5y928fG`+vOh&d?u%LFf!c57%Sif*eqOiT4;J&_gvjNDu+5_ z1J@4CBj)n@^y4F3#GWaxF5?KFrq_Pp2#z-eMEa;Zzxw&zWQyPC@Le{@Volc`oR9bG z^JDt_7KSE$f7Q+5ewcnn6WHs+g|M<1i3LNn6D*i+afj-nsy_h?)u-1(oiX$%do1D9 zD=_F7-=gmb<|r@OWh)utdHyx=4>W+$AD+iK_&g8AAaaTAe(y=j0xVAq$bcPb?)j?C zxAGW*IG8_M3+>*`Og>2IjEd9?A?3$WxCHi z9&)`_&WKHeb8((k1=+CAV2&?%*Z^{JoSo|amwDYESakx>9`+VLz%&Um4zKXl(oIE zdD{3PgiSt}W8Me$ALHmLeRuSvl4G@oXVSdu?!FT_Osttklc-GQMNT)E=avPS#{$CS z%dGiQrY#;!dHGPx!wtcXW))C-Ma=Q-Pkq}&dt`mbr`iWan<|TMNQ3Ai*9D3PT+~R4 z00g5_V`S33Rb-jPtK-@OALr(NR#~Lc7MqEg%lW#H1_rbT8K9I4X+oPRPVt@6TjFuQ za|it6WA`Wx4d(~e%wG|`uU3iN-k6-p?bbdzUJi@zv#{wFPkvz&*nWvS3k_l=Us21Y zZj@jFjhjbteFCii?*HBo?GL!KAB{~%`OSy&rXoAXEfv{hXZ~O<*^AO+JbhVx*A7DN zF9W4$b^VTaoW5*j4XIp85#6qH22=}C5%He&{EDB3a5y3~YU-wiJnRnSJX}3%W8XGBxxNC59(5~FD z&ty#>YTw5Lt*89=R{E1&aO-wri0*v(F&`LYvb4It&gfz2*u5m9F_vdlaK8+Mk$g9UB*)B)vSB1Y}^S^S57U<<8bQtYxRKk za8+0hfoMPcCgo9jV-*&BAU!W0gN^{)lA*MCwA>!t+$^`Pc8p2A>|-dt`5n(em#hs+ zRrmw$^4Dmlj7Ls*cPdhTjL{wxVQUd>G4$Ps-$@K00r*L*rTU$2f)AG=galu}E1BZG z@r+{{POs`stcebW4%^NJ=XvCxS>M0rZeHFTRvu~jzKNHzzn6+)FS1@*=#p73fB7r3 ztL1XD_MM>EuiD?x$xFj2uX~#c#`FO8WDZtTE|1BE#jp(?AG}iCw z_In4_hh051YZQxpEWM<6_TE&6%L3aKtbu2=Gh!LiGYOtMe1ie?Jh{zpYk6If!hytg z12QY5-;@nb54`(>FO!5oAPE+fT!KEA!^$E$)bTI2>>I!))MI$@lJ-EU)_VDI@*xJLUuRwJW7NN;?`UAl5mLWR}jEZLPKg!jv4zz!=yD^@W<*vR;-D7m1);F+TF) zpauv454?bEFg*o54k@TQo{cQGJR%rtWmE6g zpB0ZdOzTQlOKsHf#_&7Q$;Jf?~D5hNF-E$Ti~jXLd_4?brTBP+%B3J+uKm7H;0aF$WAvr%e3_9a?!G9a zBm+(E1CfV-Pjfy^L{RMcafBudb|a&;6kD6=Ll=(=>}4UT&m>Yv^sV(#0iQ{dx{XwF z_x(OquBf{7dqpJmi1P?GP{B|+s>IsX@M6%w%FK3rFdnsb_)(IzY>`7>3jit0|tWWFV+Xx2+ z8oeVd19Nge7i$V_u{+Sd1!TcYP>11}dVlFUi?p5>SH+Vq(h((*+m!6ERLyY<_&@#m zUnIqJKTD0kTMnWi0U{!@$ipE-C?3H{jS9}TYO8;*iixoizKn&&ptHF0WR~laUv7woWD#bHa@$)B)yJ} zf%9fL7h(N+f1NU!q6dPd7J-ZCDMOLmg5UFPu-W-twwI2Z zj%EagDoD5$p(wboj$pP-Q31j(7Ljx}(iHoG?plV27Eg`W zXYZI?;mZe9sb~?`BTS+9FBu&IxJ(>HEG&wI(RFZk)8QP=LE+Dl|0`=CVeGRn~H*D|vC((aB|8+>*~e?E>R>NdIYslL5Y2u5gPUi7TPox5_bza>)9Y5u7tVir$pXtxrLMG}4MVdch1mAl7O-a=?85?!qk>EIVk zn~GXw^AQNy(K-B368FBWfqk!B%YR?L*GZS9hu8y4D1NPAKgJu|fw8F1i(x6O5lj>} z1#EPCh1XD%*GOCYjp?$gFLnb1@tHD|{-FmOg6YT&XnD!R_}ADK9hK=!1|ok7HBpoj zoY@ZyTZYtx{0lH8N*r3aZAv&n!Ha?epV3Y^;*z(1f?LU5G_ zG0Tu*<&!%mliwQ$%bbW_wnOLVCgcFwy(@&VId0&iGc)pUJ1B3^#4eN_0><$A*D z8mG?ZF2oo1>k)J>&1oug^tWNZ#_;9OByN1)|FQ<7Gz>oOX_ypxsVACjNwScuP5natf z(F{dlpbj~a<56Xx%YyfqcmV9Xp$9Acoc$ngR6jJ)Y(;`v4;S4*s()x;;u}|^WbMx6 z0HZ&U4LvPrOg(C)g7O3$%cXd`+-$_;nHtY`YPWXuBb%7E0vlvXMsSt zzrhM}5n?~<8Y;3@zxt1~^Vn8g-Lmx4WV3QefG|yYhYa#kPY5Az5y(7!P3?1UW}`-B zH7Z7&6A{412$}V-HRm^9R6K#zlmD`;PE@f6Lqe0}j=sIEpZEoI0`G}&zMyN|3AU9Z zh-Ci|NU4f$kg0SLw{-XCJ^9HAUCFt}Rz2{(7ShwdB83^mZ5c@J`2?L4!VH;#vndl< zNK+XuvWCM<5|Z^nb#j(nR0ud)=VC&_&EO<1N$SWz_`o`gWcmlgiar2z&G>Wb3;on6 z`+11F*+!=Ou{nWF*Q6qBg0jmT0gGq!mujFrQGz=qHfE0EF?Ka&VJ*b}f=W#*rZ#!) zv>MV2D1*rT++i*Ieh&&3sdH@$s|&-R_YPq`;T)Ti*^x(`SMa4Y2VoX*n}I8qn)2fgY#_PdPK9TT3MCnd_9wNd)d-54M>h;T^Vl0{ zkMmIiXs@E3@Gnm;Iz!hGbw8d3Mt9iE7cQeaXqh>`61OwCROnihObjcR%l@^}kL{AV z@AhWBhU-M$T)X{;0WPkeYhy%o_`T<2dV|@(x%%R}#Y%a*U_NxSi-bKzd_J(iH|!H} z#)ZCq)h}->>DXO=L2E*yl{4}w*{8;8wJvKR0dEOAVJBXjuBY6dD>t$;nvK1(!a;wL7$ zsu;QpUTNlhgGHRq_$sH#E$yEhs$JQ+k4{k>$mEd`gpnrRO(Q)PtH{NtO8p{~*K+pi z`ukveJOb)Bva=lI$nLENxQ0l@chNzY5pnVmb72D3OT7D&fnZdlrj$I6g4Ys^qHK?6s9ngdSq)Up~$X>p1Gm91yl(4ndYYLA^kAfSpiS z5hrbD+J$%q{!M+W2kp(;KXm2cZG;rRpg6$g@XpR+C;|DI&-Xa$#w)Uy$)1tN za_*D42dB=7MVAV$s|hu|-@?S|AcI8WL7Q@0o(VHm*|>lTBq}^EXzytCTutwUm}kVQ zJLO*IEN30NF3a;dap>&2rnAYxcAiRfNCjZ}5In7Lns|U08IP{j<3erV6`w}t(cVx` zs?LL|_`@EyDO4^2?=To8nSwN86?p{=6?7yJ1%af$OUEaOm#{i-=XWc8$HZlWMb&-z zF1JiN-Z-2GV@h?iFR>uU@Bus*A!2@E>F~Fl-a5 zZ+m_ciwECJNBXYm(g@)UW=wG2MpRMjsIKK>xcP4uIvS_b_k-gMo?a4&PqydERvw1d z+&}15o)U|T1R~1J<)#^6I6(VRxPg(gS6NX89B8Wj z$y8Wf;nw2-8*|s{`*~Su52EC>r7sC7_knNwLb`dkA;sJ)w?uj+6nc<8c(S$~RPC_D zE}A$O-^VFWnArWCjimV6>K>Go^S}8A^f$jgdL#6aaSMFSuS1)LY3`pI!%uFDyJ~@s zMlM22)-3&@`T%kx3hv@*3FxOOqS$J5K+h_tTI+rqejb0BvV4;J=`wF0OhJ`A50avj zl}9i*KXL>a*Q3VgkiK$%F^%X_e*9{{}rqmK+X0w(1!!;{tFIyip63Uom}goX}Jhv z4xaQb-oWB4r~AYZs0kwhzjL`2mU`hiwJDW8gcKk@v$v16sD8jV@IsVY2&}>23_bx9 zGJEx^$oK5xu38|@pr?JWp4$8Eo|3kxh{nrRT-`wsbT5Jx`qwO2ULGNc^~A?198jk; zY$Of{E!8`Ib1wD}SDt7sKR=8G*?R4-OTM;6y$?65uo32oyxPa9_+!tww0FbNER=Q{ zpbkk{42taP29-egKwqH}jPnUPgD}^+uxcG5?5H=F%5KxIjeS z@7Fx^eV>&-K9BL+K=zPUvR0ZScB8D=%%Sut$A=_4aFVI z|9pbQ;qFZa23q|sE4-FQiO)#3!vtRDN8W7R-lx}TnQ!lCEcePQ$Ha0OBvkp#7d_6Q zp!up|>wy#Z>&Rb20ProjF%;!u`}hI;rumb7f?hs4-r4zc_*FNj>W0Je6YcXwH91HE z>9gVFa|g+J%w0h*Ld=a>&{q}n@%n9?xgS49HZRr7=KgYVLF z_A$6{$||VVWt-Bd)7~P668J=O%n$1T!5UASlGm;To;5)G*rxevfxsEM*XMD8*v^l6 zUG-{6Gc(#HZu<%UAR{O{R(KbCgUU(o=!gD1a`Q4`cnYicv!Ze%thnmbQh@Y(2>nB+ z9&QJoCq5R~PeiTCVW)n?+mw~_<>Ucy6ADQyTXlaY=&*8)p9^*JE@8^I;?o|f#xCco zs(y7^OFGDK00w=lb+8Za=^AqfR(26|qgJY-&f2`Udi-v}|2+(awGcVf4`)lhX^VT9 zOGrqZ!P~yOc$?x2YApI|bvA-mt3&2OKmIyL{7+NNaH*Lnd|k8b6&nUtqx}*XD)a1# z3RJi(qhTy%kqP}?u6sQPS9n3=4=v7aTI`!wLH)%VGH_oiEpg^=V;Eo@Lh34v(vwZdG$uoAG@c*0R$j;(LFFd+o+)yNRRw z+(efXq*vbEne2U}_d**e3=w&!jt=}6+_xU%J?OKoWAK4Hk6ZL<&OtX_?nybPe%6-` z8zF$4!Jt|Z1HDb_ZA~+D2?KG+uU5q;s&u~Rs+=qm=5~QVt1iTWt9owo8=!C83Bpn! zKt&*%1Qrdm9uj?5+EfJ%M`_=xGmALg1zBJ@R_r2sefWAr%fFthP=_frWM9puK#9*`ww@17T8pw>#hirQh3TyW^mkXbnS z)d-sCc?)#a@8%5W^IX|zYC7oZ@u(;EdZmPdsz&#-xY$Jx7+}TVv-wl+xCB&VOX)u5 zhPkt{5`)!h^Y@LB6{Jc#p9Mf=Y+L4DGX5JeGiPBwA*YBwiLnYpDW*@vI8YM|ZFQ(! zbOgOjo?8gcs}U7(vdKVdAB1=xP^8bEW#=q0TKmp}c+U|wNj=yd;4N=CWCX02JZARr z-NQMB5@}pmt`NPpXol#D9KtU^R4kOS^8B!;cdB4xGX$u@S6Ef06<7(d@De&d{Nc?+gkC6B zb?ha3ne->eh9|uBjDameR|$YV%v}%zFiFS~y2CW>ycwjgC{PEY<51NebB)V$$hhCpe?rg7|o!Zz`Gwip2rz zrvvox+v{%;JBDyqVHzNiXke=SyP750i^glLg~@Y5l&jp`KPQL{gV&J3GL9b-CN1n2 z136a@azAS!?BG`??JhEt7=zKl`7geabzdIXkiuzc5GtEy1Z>rDgWRL=goKZkOhz;v<%MiqSlgUf|;DMT=uyL?dnMIL<+W0>n?%5&h?#+ z@<-7Q1kyeDZy>M@csj%F@yw=?oE|3vnO{XzgCPLZqiA{hnuvy#MS*5HCd|E>ydgILyyaz0pyg2bVkHm)Q>yAGqbZF`+i5HJ^)5KcN7a&}2flEI zpx-ht5jXnHd1Wrzr${^!65F!BXjecqKTw+}5wrfa&+Mbb)`uJ;K4IPKxEXc4N&p_z z$BikwSFvPy;h|t9U+_MX-Q=%y<@yP|!+r5FvtBeIRHXRc%?ASGa~q9Gq@%gz0fAWI z^7$E-tL4H9#78IU@g++S7u1@e zA*AZ~@ZN6-cMdd2LE|Yr)dpH?IP(P

{h5H&`iC@xw zf=ru5aSspM$fhZp z9rFvi`WQQoV-8p4-J#b3%0cG3af_IPyK^iul7M_?s@mt{l*>t_3LqX7-_sR>)sJ0} zvRW8q;C|VqvxBpVGZ-@|z)6XkYxOlNV%Guf{q5M>enp|Hh26;tVp4f)y$(~%3|uhl znl;C&Z|l_~pHJUd)pF)ekqTFQFE^C!l3YX*JQ2P&Kr`;4%`M)cl56vk#V`d6kPApkr#b0u$hksv-l&AQzGip=&m?+bDQpP1`}Z_2mF;n&es9hjR;09owbb@ zeJBroX$}03-YnP(9=6lqGHl$22du5Hb8#$l#eH?@GtKkdtvnM-HOa@}Z(PY1Rioa|BWl`Nt7k3cuP z+xWY@X)o#LauT900plR;f$)5Tjs!NJZlGj=ufJSXZ^Fd&9_h-}VV^_|IX$ zJ!U-%vx8WnF`l|U#uWI^r=ScgTwUIdx@L~e=O<<}1=10oZ#Sy)OnHXqpu#L$I^_st z)p^;)`8A$L*{fUUN%Xf#d$vW;Q7~Fq6a2{tzEJP&VFHpPwu zc&AzN;P2nRr^hlWMv zwhM>~^ruR+?H7TTsb0$2^09i$$Q)%WZeuZLsEL^%fV^p~$$1|g5F0IGgT9%U?o^CRSIfy!6$stK=2;SLMtGzpFL_9B+{y3i4pKc#}+~&bsz>$ri0e1Ni?!-n#fOJ@ zb$;QfpiX4KlG{%@y(3JO1jHVGyW9wY2UX6Vc^avq{u1C;0?ia)=riB05uk?rszE1Z z!^F`71})IoJ1!S$6}Xjbva*`*HL4WJV6> z2&k~j)vbQTJ722nnNqdS4RJXEE{z(G@b~1}5=ZrWmzxgC>A2%~36hQNH@r{Bw}|qY zxf{j8 z@DmqPM$-GY@YSZ>ix~o_=CHwGpxa4XNm`qRRu9YTg`*37xfj2`?vk8fO@<%hx8Q6W zYvm;2aQSvWZ={7xqcJCiUCJrqkj)0@y1eKFZNJ^oPY1ylX75{mSLJ=r#f$52qy=M!vJJX!(oZzv z+*w(1un0;%CnLCG-jPeWI6oZeg}Pk#6(K)-PF(vm;6#A1TCRp08x-KNhoQk~d~ zHS|+J5VV%FQB7Nja~c=@OmZNftf4)^_7%{N>~J?8r_q_Q9^fWxsfBBh)j)j;G8@?T zvNsa+jkpqJ^30Wzo{?ZFN~t27`E))B&~$+9flmPTA>fMW5x~TCs*9TL-tWWGi12rV z{|b1sB%LUIJ|WifF0enX&u}%O?s#@u*UK2OiOhU95dy0tj`w}0AZaj1*YA;MyDn1Y z8>-t#?4_4ilT*G3hhQ7@c=+5*i#&77XFM?%k)4}U!lJLWuY&qb5&hduBKNZhe|s2M zuICaWWP(4b(-1jYd*}Cc!#ypAGdTtQ6NHRbtQ_o5N*Y)YqOk%z%WA{X326HaL3yoU zAM)q-tmoMIkb!>W%LkQ$rRv#>Q(g2~jlZvTzG1Lwve+)%tfUIEhoJ{Xfg&szY7SI3 zD+z~IA<`|2r05;X0hHyb9S%6}qIlNP;m4JB7I3+)?*eJy)`#xm*>1;nJ)ifGwW%`_ zOwwE<3g)who<2VoaeN*RYn1DR?|dd$NqQydi%7M8(w+ zhP$EpKyJys00F*^-P^@zzArMloqguYYzr%&zw?S%)J)Rg{X6^Lor$h}bpQD(Pac zMgRG&V^)A1u@1n&Ap<@1u~+4i z`k~O(R~V`~!yofK;ARz0Y5FX%P{0huz!qBHc{)jDPz@?F-kDe_wQQcBs`d>lFWDs@ z_>P(2$l!?bES;D39`z97sw(())p7*`dtP=%mIL4)9acg@p_f^(pHTSo0ueM|PP+Nx zUSNik4bi=Q+>{a#!`roX4VuQ%g+`#CB(uJv_MrUv@gefx(->*|o< z^l0?qpE6pPEW9KEmXR%(A=}}4^SNqo0gLQO6U8vf7+3pNnl|bW#6I2PKYOI{4M3QK zHWgPujTg^Ny}S;GEKvh zu7|GcQKNsOmLs~d4O&4@VQV0!c%(Wy zgcvg6REapuTD>gF{dhm#&f%Sk;W9|0(Z1Uy!iE`itkkL>Mg$=IvNQ!{j~Gw_0np-M zjbLrYOMSn7?y2#+ZwvzDgq(3~3aAWb;$Fz^$vy2I)jjHIK3^wxCBa(}eQjjsmxcOz)?T~QX3Ce&$6=dJj z&UxsY^{&-skSzCm$-?91^IZ(3GJR1asBr0x@0kea7NG))Y8mrLeVi=lDnAh3AfXT# zXq#pHu`)ledv-Z>MVc=-4*5G6MDIEYANOK>aE`j=nZv<_Pd+fpVG#L02;JS ztrKDy1E{lpi-QzPF03d=-uy~tqZge=L>F;6Pf4X=eQ{qwSy1yuNjjK`lj+e(0d(I1 zll0^3;D7H#{ZLMxfrj~{7HG)(%Bu;hGVZ~QoL?r7k`^5=HBY*`n;uEXtC!M(Qj^(& zXc4H>ev$T}qPxHMzbyrHr@hW!=u(3Ehi z#qd7<*)y->H0~wQSiIHxeD_oHb4Z*5D{r&e&`iOi+j@xIL2#b6Jv}S`tsxAaXE}kf z0VP5fX!y)_Z>?zN2p7SznFmo}Z|OFzWAT!N6H8|IwO1c8SnY3T{87tNSy3oLLf}h` z8_Mw58F*Ss7EgF13f7f8>O-^_XIdFAR^<)ZvJ3<^@X5=Z(3m<3%s=%!kgQF=BoBI;l`JHmTM;EJ3uc z=uI(YZog{0G(j|OHF7G-n~Wp1h$CRUQ2|v0SWnaG_IARPu(bdMA_QyD8|E@|IWz}^25 zZ?3%%cKqI(Iq0dcEqfElDu_yaFSx2cj%|MTLjoTSb_S;+3O2Xh!F>)~DHPEu3~8GP zv=hmNuhk>i1N1}zN@hNMeX_wmqK(b?EN_6-(C|nUc+`F;V|5_u-f z)?q{HtQWrI7SK3jVe^5{K4j(5R>XLNKz{53lu(Aci(xC zs3?w)L4=4&W+pdylM9;Vj5Myc35EJG=@}QM@I!yoV?%!&P)RV-JAX-%;=h`%ng@ZY z!J@0s;Xyv}ao@E$XT}T~3o~>?*A}@NpSs;!&OijyE9OuTQ{Qqm*pDyd5-0^=(uATO zAQ8U>S9SkTwgFZxeX?bZGFfR+pLSq>RfA%69~avQ3+=}(=nf{$TLR$Iw!rhkSrYq` zu=KHYh(9Y(kOA1f@Ta}LDWW@)*COt3;bxu3>vIEeAv^iM{Q9y982V}122tUcea&cE z6AoX*{xKsnyVh*|B_V;P-6`Tkd6-zll!id8ae(ukjz;32j=0Q#nITF@C;d8)4;7v3 zNuu9R;a(gUeG%NmCLZ&Gr)&+6$umHa>EmJE1Yrp-MN$QK+qAM&NKncO?gnuE5E748 z8+fy#8S|AON(ZCXR;u0tTG{3mbex7^E`~ToE|#NV(I}wC)KN!x_IfwE!(l#iEEuA> z{RLt>)*rCb<|C6c;c<=p@3}+;YA|4?-wmdCe@^Iz!_|pb?~WAyKMJq)&G8>dRBA#% zCMt+aK)Zu=Gv~F!W~OH7$&=k1?Xq$&__?9K2sxK9RU9vObXk}fu&;zEu-;sB4gx}c zkq~}VKlh3=V5!pymhrlSzKgXMocx1J72PFmky6J6z4i8G z>{k#OeCZO?R2qE$(6mdV_27b|{y2Ic$yoVqKC5ApO0}L%N;eASzD7W%uT>T@iUlBR2s$FbK1$LgyMet_!JhY3b>0&dQ}{W=Ux?)X5}^xWqC^B(Ne@#)n0ST`xOU2x)QebJCJf_Ayp$A@{ICA*Pl|9+w_ z0Niz{Cp?CYd$=Ri7?`PzqD4aww>NxB;g|DSs=zAD6p*fzn!OLSO;hR#o$s%3>98O} z(7tQ*VHan7eS4{5{-jr+7hP(&_fkWYbEAu=@(vR#Q_KVk;}^o`Kl|?btHn5M)6B#Iz}Jry4QC&KW!Ay2|jgk)jO22{Z*XzWEsv3L!$BBIN4~;TPQ@ zcL#?EB!2XvJI!BE={i^`M{`!ntUBYQ0n)xj_K1x^-sM9X$TII(+Au^ z^g~pp|NFlJ~=86LJGGBQmmpi~g8m%Au9l$U;<%s{h)a z^%h-@vq9_#B@+i;LR0K%^#;_qhZPiWax=TvVjrswIe7U#vE7AijOYqA*yAH{9YplP zO3A2I-N9(!Vk^c>%tfWTy&C94#i6a0HHEGtT2&2v2N8Y>GnmGHg>CyOCpHKBIQ+7U zP=1^z%G6163;)3aOkxWRc<+uAY&Yk$Y>0=Eedu8n_CWE*@tTVK;=N?%R%7ZL)!jN^ zA#^oL&zN4fB$-fXSgePTo^^gBsdCUXEI1Ju7~Y3Zz-9Y~QGf3BLrED3Y| zMTi2Kj#vW6zb+j81EQ;38i=6m*P=SUI{TNL-kxEzvCz25Nq$Si2x$ff75RoL3ZNkr zo|)!L`mKa|^MX@qIPeMxxm(N*?^*wStT!!ef2J1i(c(vz{U+^>{!cEaF?Wlk_Op_M6c8A=t;wK$5zxFY zWWT0CJnJNl-ZGvYOPqv-aW9V4U9Js>9UDaJMY3+VsU25Ddkmcz{;plxD!&bVdHnnqEk|GM;(b}vHVL&z8SwVF1Fdx9kSX5}1LJllJTGy53lx#lUL86q zD!~`uOzz(LE+9d_SYYAw3qLyh0iPhA8k58m;>38rGUi1?3B>?z_27uG!3guE|DvgQ zxf3J|=fMru0Sj)@^roZo+tU_>;_Fxt+~p$OTlt1HBfD5HaJ!D7+LpGl*&o@ZS)~3l zD=6aa8C0>sP6;md8zHPlHm3*QXMKg);75_h--s*2iBL!g;nYiu_MV;od6vFNMMP8n ztKPcahS6rF`}>6f`xLna!vL6>yd=fK)yT9ZvqX2xZPA8 zJRG3-d73bI2d*}JEPUATcQ4DARqky={LIaHT1Z>P&Wm)pSNjZl+#a{C;dv&waSa>< z`xzX+>A%=y+Bm(36zxoBG44DjVk5nmiLzn#moJbL=UP?Eq`n5y5?hF*p}uFd-G%5s zj(9b6IEf!}iPPmjcFXMn8+^a1k4LPB|GDWnWbLgYdVxy`h%alY{9oh$`g`1w{|^GQ zD4c8l@4t^opAyrqNvCnep$y`=8(U5hEg@S=h9$|DTWi z*TW7(@%o=O*5#jf^owi6QT=V_e{T3pYyGbqK6%{le=V9g!=KFlr4&hA-{U8;i@HxS z6+AVMDgxs~|$E&Ef&Q~V!CH2U{PiUqsQ(VsiVR5m|fUCHX+=;+6IKn50IE3@)P zUgkvQk{&T{Sd}~oX-z28$ySp+ZHNi|_p?6`NAQ^OOqtGE*(zIeuA#&K|M%bi>@!VU zEV9jLgXU;krs*89APc`Am)8d^;cgl;M;bF1{D*)1`xnpBGaSQN0Bx`_udoIwXR=4TyA0{X$zxbLKEfAM$sjSo8fePA(N#}y6`-Mr74pusBvF;Vrhle@v^|HR`-+`&`vk^M|)v=x7a+abb|+i`EOexvnT5T zmOKDL3m=#cOm!sLJk%Pb<$3iB>3Mw_iCz;O@;+{L0`}q~mXznLGx$7=QC?u)3Hxmw z#Q|3HaW@#CXU4FDA(VPFhzXqk@4Y$CACq5_r?hIVN=&9pzMjEBZ{yP!X3_N7Lb*Ns zA6x0xVF%N2#K;ytBW=#y=Bs*(U^u^&`X#b{Z^3#@lK8GcW%8Ej!x`~jG`LN8P_K11 z(A2a2#C;(Ysi2Hj`}Itkg6TZzrAjjo}q5^{HlZzB?^@CapEPII$LWWQuzCPYNFF%l$d?5LO{`Whq>$mro@@x9t zEBN>zPrVQtsWN}`fyuk92UMqL6PZxM;iqSg3_im!Pu|qDU;3U@!|_QBMl@aO>tJha zbcnuA({cXziY*gV^ZxMpy`}gwK_UETzIj9FJ*gDnq<9N*wjsrL%;^@xT2kG>)C{>U zEKr}25`^K%P>QGXEBYnI$4yeIF=+MnAl~yWJX_CEF$xD)=Z`Nw`Wr$W02mZL`-Vk# zQQ+!hlJPK!gBm=g(j2`Pn|&;@ho)T~yqnwOm36m)OO%OA(xaq@#_)vCt=y{xqzR=T zPNy(c{Bj7MwNkW_VWqOa{tb+|iF1Wlfx-hz<@%fx%!SjPi}M%{v5_L@#`foMF~)0B zR($bE(=A1u*nzmAa1D}Cm(hp~YOC#LJ^FUOFb&E^?TIp&n)lRLCU>d%4u?`#nV;gm z;*Fz+q;s0@i9}55U80G4QX?qy@^gFkmc2xn*>i$M0t00-Mo%K58s=7lP3E0a&~mPV zThc2)Egq?BUoagu%POBFeW!(A2Lb9Mwq8a<`$pt73X@5?$=ZQ=tA@e10kO3_<+p9+X_2FADsy37)|>_dHRbb z`R-67T&e+rGxSn_E$*RJ9Nq*VaGw_?_k|K17Jmv4-^CG!$ov{pWH?b=M1)^n$Z*oP zgLu@vND)7gi0j0B<~`#2YTwTvg0@f=FhF*+4v2E~x2uokL=MLD8?3jo#%7v4Qhw0+ zoVLOfT&XgDU90|xLRNSQv9m1;h{A4oGitZ}d9gbv7@N95@aQ4w^SPd|$_UUb7AS&b z9NHYJqfJ2t$6N~tV64C%k%6=8VGx^4^uPLWQbf~k#b?2HJy_;;a$5ZS8Qa(4(#mCU z1cmq`wufmBUUI@qd3hY>y4VXsa#Wdv>p8;>_7-FBRD6_%Dr=~ix5YUU5QQ$}`d8(v;@n)DHyk!$sG?BYk!Ukf_>fhNzCx;=&*3h!saCQ7^ z^}wAfB@z15Xj8SnIK`uRq`fJRh*q)>ch7n`xIoI(+Ix>QNpJ3D7id*egHy--RUu6V zhT!qOc+aD_ys=+1;q80#qniyMvRl9Ab3#@G5i4UC9r`m$|76_PcKaa<8#1CHV;_jW z5>n+#`n5klcpc5=w-F|&g~*7AmJfpM`IpuHoHIL($;G3tz|r5%*UvT7&gV^D4oGry z%da{e-^JO=ynL5LGEk`@Vrt-PR9OM@e0`6v^8tlA%)ZnFPkBxPlvC)3D!i~b<*1it z`;F;!!cMvHdGM;s>V77vVhhpOui}bq)8*Gv5aXxa2pQIQ?BFlt+rPlD?!KA3W z^2?|_+e%<`Ox~?q%t!rKO{{HRu;&Px+CqP2ivZCj`tfOtM{x~H%I=pfs9{A|VjyUe zhKkVJ*2weixm^=j-n3M|BU9G>_IzUFo{RZ1 z-FrxnPvY{(W%@Q#?u(!)5F1pjg6F2!kGLW%>T(Qc^>@_!lnQS1CAy3o1SHQfME;r} zZ`&6~6Vfm7F{4Z9+s~q#U#_>ZaT`TkAu^e~kAmrAGvo14(j;XrqIc z?R|h63&g1D97V&{TVa|K#e2_kVX#(nk6&Y%_w&tNS1F{T%eQ!qxQskBP(uXlvvQdn z6z3}H#PvbP=OL-XxfVCL8d8* z9oy^1`0d!{q;3$~Mz$IaVW~VKduK{5{^YvMv7pFUY-?0G)@p;<$d3Jcd$@2ar!ww4&BA4T|99J zGZ%AKiI+yKcM9Zg+I;TJLZ*IMjHG;MH$$8*ML9;wKg*XvJRMN^=-X>aXtGI4rhrZ{ouiUu*8lJ(GTkH&zoCp^pZXs15m!-qjj zIA{t2{x#0pUq9FvLc9q0&@UN9Gr0{Tvd<|>PaTk*UGtR>8>yM~wI7=r9Lg9grGZP| z-tfAO;qA&1LX_b+2sVWALZLia=u#p!7Ie%=CLFyTJ=4952huozWy>6*RqvOKl&Iqm z@#q&)i%I+b6OaT6N?NR)kJmdIScj`ZjWoq1#Nk>448>Oa!cOL%U)D?~Wo(7e_`T2m zP=}vSe}d`hw+~PUyMpHFt)0$a|LA}Phg8PFwKDk=qND9(?PPb(+O|pZx<*6EF@twr zGB04c{DYl@j6ia!HLY&zWyLx{0Gg*BAJ@{SFs4B-CdO%0sgM=>YS`Cr@*uc6S^J?j zQIGBgsMG}+Y(az!XBhSTPw($1`M&4$H>zgEj%w+;S4VdLT@H`srhFJ4VDC10?WQj{ zNSKk`rfuR4qzC5yk)6;3{l;_T!0}82cVkA!{dE17>!|{MQW34}UBqn@`qp||Lgcca zVEQoo;C`NZsaA=CKARDcEx+hSn?1%pK2_nF*yA#sU`9+&b zDOgKz;;AQ9WYMK7Rsq^?y)=-$8rn6e`#%RG0d{monL|uP-|5w!s%gDdpH4c8^LrUf zuuRf1L)74XHI2T4cCnFARWp(!92unDOOF#bgt6X#gDi#GgL;cw8iHuSh&yWQ0O>e1H zCXVuhrs*u*6K1;~_By@_O4T{`rZx-^OTB|S$uBM3cjKl8n~Gfvk!*!suPnuG0GRy- zS<-zOHI{qwkV91(Ne71%KL#3)E;b;7Sr47z|Xq>Q&+Fd zMydo$TX*Q`CF_L#k=c!Zr}CNHj4r6ap;{k~8uqo~q5KfW0LYtIiL>gRn(0sb6>X^8 z;5>Zf%eRv16@9Do-sr=AM;u9S?$ry46*hzu4Fz{P?ul;`13-k)I-n#xiKFZPD4%ye z^v_3L=7mnDM4bA>(C0E7|FSEXOiYPS55CD~c^W4GMoojYkJ@Q{Mx<+W8-7B0S5CLD zI)BADH$dyN3yK$-I^Hvmgi8iB9llKqid`Z<-EV;Q$L|QN45r^GG=`XczJo*90WIc0 zMw56?7pj1oi#z#-~L;lNoZy3!JAwb5u}bUP82{xo5OS)|Y-N$TVS zl8|oob>G2`5+zZj4j8%Q$SsO~VuM$5rPZ_Ag!4=Pa-_10Kqj}pKkRc9626L{`^VFx zFYIRshM2gZ0wbHX_=@%C?Gqj1?bPb_nW)(Dw^tBQ@9{8hO{yV7IG1FD3%bp4pE4?A zgt<-)0?#jw_?4-zQ+{wDkeOOHCv!*z4CV>Xl0F}A;oVayOOKKmjj!w#U-tHW>{rbn zmlOp_q9e^t({&!VFLm^Ss+6{)ix&9)5n)!r)UrCfjk(-(xzO@EPn=Be`LZv=Re>Sn zof3O!og#F$`)fS!NHqAlnK=jYGs&OVL%SbojwhS$)uBSvx|X3=&$RKMN6aH_P6^4g z5T2jU^P3gC$96l&WBX=#Xru30S-n;DoUPuCsZu68y?wl9x?(e1ZCTpsJ^q0eCwA%ll|uLG7bx7(}O zHqc{2lLiqDcC7*Yr4$iZ_&lyGM$kiM^U@3zAsmN##uD&~)G%q^m*)VnYAPA(?)Q%7 zf=4GSqOemzc*Z6d&hO}AJ%;vj!HNP`IdLjNqK0)%=eTwFu2^K9ndq$k8fRtCUpyZ+ z-)(vKXEJSnQgv==d_1Ku(tc3-`0D>)i;D0t>!Krav#-V#T9bCauZ-=Fz=I01P-%4} z*#IWG+j+9|nx4|4*t)4_yFixh)gkKBH%#8xjAM6O@u{8aj;J?Fv~b#L;1@8cRPN7)am z+c|s!^c<o1<$hQ&-@<~DqpxD3w~2aox8wyLCK0$kW)%9IYFF~C)~oc;xY=rcuQlDTwU^hv z!p7DYlonsT(A>C#_nMyC&rHG#iAIZUEv~i?3V%rGNVs1xel5g%0mf=OfwD`HbqO6N zs(DBzO&js?(wu3x)RTIso5P6M@ZVGvl1b|6yP)nEiNdZ65cF{pg>g z&kaW0R5jMeV?iP-XOg$`etCQatABS|FS?bPX4E?chM|(b?|t}2?-Y%ck#RRdCW>Sw zh}HA3eH;!1{k|Fc8W*GzY{jPDcPYZtNKgAr_Y27q*s!l2Ki||NNR>`yD^a8s1HUDt zq9`!hSRaybPf(zM4N&TCXl_MbJ_-Q}?e1W0+n^r$`_m2=zt}esUkEO>il}?uNOy| zdUw4B%8mO__fLxR_c({99^H?4J7_-KH;?OyhB$g4 z`NF;M{Syncz)_xa6wvg!BJI*Ze6!QL>HYcgl=OxWVl8phR)Q#@7fwm;X4}T&ZLL=% z5XWy++=mBOew~U8&o9Ho_=>?o`xTR}n+gUZBd>0`l99%?%ptsbLAfsUXyy6*bCfZC zbuFomX8~;hhm$Su0EB{Qa~_CsZZ~NMcO1@QV>b=lq=pLdm1k*`Cpu|_&0Fp$90r)W zYP)VPwK?)T^5NbkH~ZIP0IP3)i6t-{uX3$W1N6E_AymWD)zr_g{1$Jylk~l1oqQAd z*M8ZJn};(Szd9>p9VG0l2-B^Yc<0a6`h6TcJ(ha{MxvECJ5~q{jCg?U#5geBqRm!QXl%+fhndLh2a5KNY|C=YSmcOgW~$ zC!MZuax^B_4L@HbM>N&9Kg|T9YWV3V9=uW{By#9?@sbhe2%^h;WUi9Y7;olXyvIJn ztM7&8@jVMZ!sdv`4zzb?k8?L@`#U8_OMD-Pl0u_)-}jE_rJGOr*ThZp+;7JT`hRnG zX3eT{>lOY*Nr8ySB|S(dy?BS9fPnNO{Q3;mm!wXvbCIf~cFwoA%cTf7=i7{TJkLk? z+%AYKpU|_#Qy6LW1}|Af*@X^izZ9X`2g@bxl zeSmV*pXR%nV^p}3(+q#-^<6ddXs=gp6?TJ+A);yf+Mmy)I`YCB+g#UOuL_Uq9>6VL zy07gbdTQJ#q%jL4jO!IhbFI#!8+i)Hy;zJ{g`0%Ie{I^ItknYAobv9D$uQU1t4zK1bt;z+Zre3we*pTQOzxSU1*51hEL~=tVoQPvOU)WV*JHB zaC%)%Z}M^OE3*n5%3cSiI`)G#z+MRNyCBb?SxEr_vRjGr{r=g2PbGGdN%phrMuME~ z-M6Xsi$JC1MpDPBt^mIG*B^cz26Tf~eX8d5(in)$Dy-jU*^WjkKBG2t|K1x6@8O9T zC^QcId++&j2DRVmiy+5zJf?^lT;iyf^GOE9CBh-pcJRr@XE)$9$wxhrRm}NaOCfXa zw@Z;<7!C;$j_1n_zkN7|-=lGuPEDU9#)*0FrdG^2fdIJg5x3zNAwV}&2aAQdVeX=Z zyOai&iQFaf=MusKrU$M9jM+zI$)tKcI;+djX(oD+HJ0wJ|TtZBp+o4Ua0*h=;Ht z=;JwmH~W^GK;VAe>uNTN;?eV?eB~apb&8;cNaK1gXU3TB+2k^6FoT;e5k&gr2rnVJ zkAS~*_C^d^h7=ci(WA__BXE~qN!!eo|J*)yM#~*4W?waNHo={j;6Pvh$=Jtv(#w!D2renD2(*9=%s>Lb7YgJ~Rc z%+Y0~$-?ItQer)B=7pkd z6%W^{{zDbc)~U=O>6X)gSFKd;G3(eb@$<3|j-;Hp^mxZ?%+=8ztA5XjM^T(|G9djR zoQC@sX#^i59UI<3U6po_)z=YopWfi}<|GEF>xVBkE^Kgpe! zUU`g*%gd(ZkvRu#nk+``=H3O>3;1)ENm<_k7C`WvsY739+YQi+Y`F&b9dELCeG9Lg z5i6`_^akY=BWCAiM;rENaJnE>FoSW{pqc8mKF9!lX6fka`PrVm1?MpOq{memOpgr- za6axnp;=ZnpWlO*5@erlzOUR+UbRyaQ&D&8)^%rkxX$-BCJjvN%$DQT%uZVR`N4Dd z-`O{FkX%U36hB?$;a|hAIP+epdOdn(MeQfEdvkgavm5h3d-DMU)LlfV6ub*hVly6I zH{l0+(QDR&xm?j5C-^r2$(DD|dXK{ad2^_yIgV&Lz_^SHub8Py78}jXp`Xm@z$q?_ zTFr1oA<4d!-^*3?#y!ids$ae@8t+@=Mi2WdW8&JqU$8u3n6G~88>yo721NAtu|2$3 zedSq%3pQ}q4|s@U?+o9Vx1<7p|BMk96<&7T=^0^=Nx+RQQ;2mWQh8!Na9~s^2!fC) z;2w;5Y66|KvV@P~36<-L9G}W0*SL1_jk)^qELs}3y{M4+oY8|{L5x1`7g*dE)B-hE zycKB$?>M;&)Y@Sh0&OO&pqM;?+_3QUOExB)-g3via=`e8P7*JV!C({$it?Ph;dWs0 zQqKV((%^2K@60)**=R_~f-C0@`WsdK-kQ(*fCB2ezP{R1XVv{HmU%qE>c{u*9#2N{ zeGjM)JV+Zz>Ze8pZdmH=2;LO^{4sE7{jEAIn_2|PdWP1Zm6HO!a|S_x)YOiYZ!i2E z8b~w6ycYgC-Tf+*o}f_J(uB>y`H=moR=8zfu``g1o)PTR#~Ve%8p`$bJ>+Zrq3W)a;j{G+MMR_ip-r{Uq zdmF}`S3#Mgjj-KSSS#V{V;E`_;w4v)VfH3;AOJHJ76KqrG6IQ_yAKTOMUYIGn?Dt* zoGI{-Nn;hGW$Ildl!X4+^i_uoBB?;LO=9=}7(cEf=b$hcIqG-Z8p~ zZcmZ&efk|AK#)Pr|D>6aINC^x7^Q-%h$-^y$;e+CDa$aGBrmj6%l4uAo{D$W56 z4eqxOC>zO0dSinyyNDkP&m1wI$0n?U_Z7o&jAn`?*7;Tz2sJ+RtGU{*6@9Gz7L$ap zdo3Q>)4#r)eavwQYi=K1vCTKkgoW?r{8q)tyJphYS$X(b!zAzdz3pkwcm?H_t$pF26tYV!luDuy)qxrP#$(?0bYe z?jgX0W-Kad5RKkfeOzwN)xdZ%en7bED;fBv$UsKL+sn^e2lwdyhgsj|d|fd$d(zTy zAwg~ejx8WrZ9m9*xmB)yMq8HFnX@R+m3_EguCRJMy%B)C7ZP-MmVhTy;@Jzz{>)uI z8@HG?9hF3uftOI6x002p3>kwzb<#D_Xl6kbnf;u6PE-bap{{SZG4B!HNW>baP%J!y zCg~}$ZgKc^axG)Ua+gE;aJSzFrm$xRCESP= zBW*e>crFH?#e70v^e1QV6o4PtC+AqiZBX%IybnVbk$mBEw91A-w=Lpp5Kk$cT=xB)(6GuLE~56|0z+_(M*yz_jBs+!0{MkdI!dS4cH}^ z0OQ-m_?szf-yOWJ4tz>+E|2|s=cvKpK@rCO`*KZpzbW%N)(uj4a$~*iQ>RB)iPnbZ z6Mfsnbt(4SyRcw})|U_WCG?&@kt0N%e4b{h#FCC2^T?wpUQJW)ABJ&AX~lk$k7*im zN-rO^f3+F7P~n-cZv@Lsb{o4D`$Xna+F&Ubbsw4ROUMsA;{!5`8*0DrjeMxy7eNZ? z*{4}(UHGHp=V0Am$pM=+4xj%W5_~8b9+ME6?p@6f{3X$4v<|HJnjW6zTut@DsQW4< zCxRtU+FPIzM?oA;ilTJ8_~>#{-rgC`_#9|TqEJ=5F18hy|G z@e-FlW|LsYh-y(c_d~;!&h!?>Vo8n-dse7~Y7CcrQdgUuEv$KH?s)iZjNZU)P0a7A z_`0bMnNFC(|G|AtQ#9c{Y;dd%w?wbyM9CgcnSAR#i*Xv;(P1eUDA(q~D}uY~-(gy?yz$Qu=V7uSekIuWdqg z%oED^Rclp0L&$lh2;(q&0-Wzz=QTJ8>qaAejE_Hn*rHipd0JkzXuOAdl)~b-*2_9T z0_bq!+wv6pl47V4Z)FRZ;xO(r4l{#F>|9nN$&mK6pcRfP(+QJPVSa%)g^3Jx_V-UD zbZ+JuFdBL|Ps%F++{A+i2jbwrKe+z5W&SS&V*dSrWM+T#Ulrs3^6>wOV%$0X8>rDDGkBPWwj#?C@5H|m|9N8n{|d$>&f|FK ztM^I<{^UHqc^BCU-D~h}^GJ^KunE`n01XakXvSMwG=cZ-xA>%q(rRqd!PqJSkGWYS zkIBuYHIc|Mv^Hw8H9Zb2Bo+01UpW?HYaQ>vj-0b_8lMA6k_Q?@>8jvV8Y~zC5mItt zSwO)tuQCro`tO)2@c4q3d%mn3P+KORjOE)~7SneKi$8yW=R1OmGn?D)diu|+OAw`3 zCSECqay>S|Yt#JyttegFU;PXzG+3uwe(52rF8@Ts4!i*oaRYY8@2Yf3{u8DDm(Kd{ zzYwbDrUc7-IFbE56K40jsQOQ6od*(^qi+T=y5InO;+T@=OCSfxE{3pv(z=V$)rEgU zjGo~aXZ5F%nkxGi?6EvSMtIgv8Ucayx=XD0qjWfB*`?f&HUAHBJPrn~ZJe_U?A@&X;X-!=Z$D;fohhBKh=APZ)ef0N_ zgh<=O8QEp}B3==Xa90RU2spB#;=j_zG)kotnOgp{k>rp=oh3tlqeSVl?~xfi??;4j zR&o+wLO$xsv@GP0G@&21QfJ;8IgxUaIvw4qA1%{x`i$y*YDzOIfk)Rgl(5(?y8A(& zObw=jS*kAZ+{h4xwkV540TyfMYaducNWtlvw`8bQohHPGf>1Ix9l~IAsowK zqWyS1Zt8_k?{0GsxyKW)gJD?eAh~L--Jdc*s(uK`{}IBx!GrZ!+Hf|k{o6y(9+Zx= zuIl;=kneikWygv@xxLWrd9OU)Gl6J1$sAFtNy3+h+2C*ldz-j-oq^hK8Q#PuhMLDZ zj1KH0qnAOd)pg4&?UdVODC*}G+E||3>`2XN%+{?4U1_0^(EI`8tagq2L&=l+DzIi0 zOM2{r7l#>HIF;_pjk!xd`7Ao04#MI4kkDaiW^fkd|x76Yc zaJGvtz>I7fsxf&Fh|GK*RB#odZvrKgzKQYmp&pxM8S)ukKJki3O%5Dz&M{I94rj)v zR?_X;cW{9HKxzkQ9zoJC+Ms*v2h0z*IZq0T`0Z8Na5aJI_AXqhz!e9+9;G{eBFP2J1p&Lh1D=-f@Ih6sa%= zxYfb<2BXc-yS0f{Pfi_}9)Nvu`OkYD;6m#}*Qi4a4oet6s}~y|={lK+EJRd$mw9tP zxAQM9wupQ_DDf07X5qZKH}B>(c)D}U-L30&KuiA8peiHxsLHeOE?6&4pI_qd7K=!i z_zt!@+rrHXh@n-F=OqTkDKJ|kR<|O09#)X%a%T2R+fP*wtsS1zz{RM7w7=CDL&trE z89GE`e0Dl&r5%@T5|QISmQy}IiO?u!w;7_v86#UeD^f_dVg; zhmDWfSGa*(J{P_8&+VSTQ{NkpR2q4qUfvy;M}^!y$K;P@nBt#*`YkuTS3jRtkX~Vg z*L%$Ze~xXV(hoNa4!*?qE6OVM4KK~{Afr~p=8znW&EXij40Lv7ma9^LC=f)|FFm7m zSZhMajgX3>2Bo|rHy}TN3{^(6<@W7bJur7{Ndns33h2KThJN_{gs|+%E;fOamKDGj zq2w3{`SCa=Yr7xLBY%4mH$aB8^v-pZGFg53MZgLOa!I<<8oO_+LU>V{E31oP{{g@&~@%ightFmD#wR7Awl-5BD>fgE5RE?m63!9x?16Tl^ zuOo<>;eUOD8vtHJUyp4l1s;5>;YQ7c6g&CAl*o5p?@$Cu zp^)A{)?x~8NkjEU0VV}s0$+;qV8t~?wuDM!(zgm{gWTR*U6*$%{!luIR}h|}%7wUOS%SBX=C(7I_EC_n@5&8!B=r{d2Y z$0^@^NEHrsw^$&`@GY877kfJYfwt#%I<%xI5HR%c2>q4tv{{U1w-3iHfJF)(}5!D(2}z^1c`sN6+j% zVbisKAX&F$4440ut_QPSucB*}#;W^J8gjBQlZ8U+HD?*%vI@x6*`)Wx`12b00A0+u=+r!JTF9GgGU1Bul z6>{ZSweXf;ia3fDrGZHFC|333fjOR2hjd8bVnw*`mz5QKFm_0vsh?_T(QH|lun}%> zoiaInrj8we^UfAMFSMUh?pk&Bn{VhUbb3%_=_XzLuE zey`UVz3BE)Lc!=(GzLh@isUbvX2WrND8E0n+Z#P}kl4!xT~_v0-M-)M#U1Wt2bGRr z`*R5VasSzu2~!o+@%3H~gbU4Z&?Wi9M(-OxxgO!{#Zk$@fRSZBG+TsqiSH(aiE_8u zfGWK7)(}w&{=N{gte71QL8sU7IrN&oarfV2VVUIkDdBclvduITJr}9hJ>t)k4}b+6 zy-`Bj{yg2>#*Zi*KwPFng8yIhZoW z?BT-g;}3rK$L`Rp=@#OjuyK;MjM0PaJcFfszzTq!FfCyRC?#I@Z*MBMfbV=_Mxnz@DvE z(&u~i7<71;7ytr11{a0rpGTtXs}dIiCek8V%!u6P(AGUNDgDeu2Twj*aUn@xTna~Q zUomEMRopRg3h>Dfo~|@l?#~!ew4)<&)AxF2@v~4t^s42UTB1Hj>GY5-f&Q`eNrlG{ zCcu(%iVc=1Fxh1_^b0Nxyk)!%Sn<8+P)fHN!&dB&e&kW0QBj8t8lYdyHwa(rmxR?5 zhMIZZ(+0fklqbyFNnYDaIMny1Y1`R3>IFs*`-=zD45bJibLmK8z6z)QUKD!)cLV!9 zK)4Oi0aOw=8(PJRpW{a=1j8x2=L*H2f1lLDcPKA=5N;ZQPKR7sI;6|&ThDM*=|_BB zyKC$Qy(;Ki_0~O)8*5l$Md^*FvzOqrp}q;{UnqNDyh!M#I(eW5vKPJDGBNBn&gL;Q z%)B*Tr5c2&f#-FeftM2SPOul`6e*ZZ9HG|ci+=)i03ZvpR6mP&uboc%9p_z zBenjzG~R{|NjJ*T6U5(tfouR;Y+#ia{0NUmEm@u#zIbujdG9NxnBDGECI*!Y#Enhz<$sR{|T9Xl(%d z(fU4MOrapD#dBXmFu<-Qr5=j0D@#-~{P4O25E< ziPW{d8!8xg`0j~g4&7aO3^LNHN?jcE>2`J=so5(5g)1$-ch8`*mx_-MuhE)Ggl* zTO6aFPJF$Vt&XSd@+_uS98=_A2!6Vdg8AS!;F;Fhj)ETYPS~vX-`j8B^j?)rFuqa?o(3Q9hMqgwTe5Ia~qt}Z`Q|SZya1y5N2ugoWtjHJbvQ4%ooo1n;z72V<^O^l5Wfa zS~HY-gFws@i`uzsPpGs$(5GMNI3C*S!)O6K5fL*AWd}5gqtEDwe6wMdoem6D&GhyH z%LjyMTUnFp@OPqYpfl!+P#s~zqxbsj1D`8w$x%a{>k(O2=*qJ9Zi0la2L0PU9 z>SuW>k^u7z4xo>hHWHJkLIMo*GlXMPgD~us4$@He=Wvv? zc81`u7~U~> zaqe=nynSQ8_6vIK-de`3p!d9}U9{oeP~8LEv~RT37=D2jpAdv~RatwKNrnlCtWSQ? zK=?*O{1U5S|GrZ!d9p`03%T#S4C~{NDN3Vxpf3m0BEZ&JdY(wwF8^RP>*kepgGmBi z(YYI7xuWWcd4BzIiTIeo1P5vwm-yjIa7$baj_UUP1aA#jvKqxG*sz4~uY7Rz;d-T} z06#2%>sZCcH1yhz>+(lS2lkaQpnm(tC0MWk00M=Hyw@I@DNazlcV-gibOVJ{dt<($ zlutH_+BGFtVF8naCu{r1j^TG%r7uAyEn}(Yk#-4t3FM#g))Ptoh#0IXR2~MLaG_*+ z_5IkAFS7EIP&M-h9Vhio;0}1&77|eN09D;AD!l!sf$7^XH-<5MxNhN+&;)=GSr#t* zT~+3U73d~cE=X=FjB(y{c|_14i;MOx^|WZ6N*5w=^j{lTi#Pcxb$6ceX^5#VIJfW9 zGyi4a76RQ`*!9n`S_u4ravrZ2BR824K++T$R<1Xm6*jh{i)V`mCX-r==X;?QW9W6NtsO*do z9J9A}PJ&c?yC3%LKWY@>UOd@UhfRG}KFDwJurL`c9nK??mlf1>`!T%-PRcW#{2l6t z1w+*YSof70QIq3+#QnTel$;x$j?dyj2XX|@_-xp3x}kUz%He0vHpDi;4lRcbhZRLY zXsY7s00m=h;Wp!C5LE#)v(_{w=|EI(G)hFGauR)EjsRlaRG4(aNxV}5xJn1<>lS@# z3ysXxO)n12`E-nb9d+xx&-YWfPR5DOUJv+jy=#}2wB4O#&o{#5*U`~qC4PE+zg(iR z-SfdtWdT?^_PKO_&=2g`@HmQV0S<=QunA8e2xs=^vp<6W~Lc)3=1COc*;s>xur(Li;@7jPqi-!dMsfgcyRR2aNFTreQo) zw5QWS&z_~CcQChxGLSGT%RADTlG=DUcHbqx82dQ+o+~UW@+vTvt37rJtZb$(Ov!aH z2t3@#4t3qX2lg~x-)kSj%{92(-Mi6S;C=0VkLyM^uVfvU1 zUdSJ1t0USG5ppiCKZ3Jvc<^&LFo8!0ZBD~!T=?9V>ozbCLD~Z$ZE_hh7rPC}foG&b z8Ut?obqm86^%u#lLmGDQbXY6dFY-jHUjEG&x}d9A3am-1eNCt(gb4|zaVA*F_lWu9 z)C}^xU?Sz?f=UUGE{Ek2y6L^CGH5-pSLR%sa=yUQziW?wF2(S)sFRtY@}w*X0ueE< z80Y%p2j`-VZ5Q+eCWw;{F_8AfF`Favbv?l#ZO{UdoTZq zES{A1BI$Z`rmMTgWiU0H?jc=5QIAB@K}*-Kj@pvn!$_7J{xGNqY7WNu||?a8)xb3tqOFZ+noqR=76p_Bf(X3oE*z z`y%^zcU`pafR7n#J&vF-Mvn8`_*$X?xHpJmvSokFY&}%fmXMwC&FD@@(eXQ7QCu$R z7%JW6YMhQ(qw(M*^9To=p4m6Z?S>4Jx1~g{D&AYR(aK)v-JgVC9G~8Z1EdQ)OLJg} zq?S9KzO*nKsq*kf7?ua_V^8d!MEv>OxZ+(az}w{b4)VC$1pnTfKSW8~Y?X8jy1d20 zeYi}SnTDS*3VLw!y6tgCP)T&MlW&%$|5)m z^~TlrVu%L1szS6Tna4KWcz#F}_UNLoAT#stMbYQO50@Ivk36U?o4q5Lk>h>E{s`{a zm0UMQTSBr28zZLR+##`fsODpOAxutDuaH-NLQq&rUCATsPHiK3%EEIpc2*J$8o~gu zknnTZ!y~FV0n;pkOe8Om$Ftx<{WvxABJngcXBLYXnzeT+61=fK)s84^*Avhvb+pc! zms-lLq-hZ zpYQjsxkRJ(UMl_NP{Affjk}iFBRU5M@KuEyywh87VAp-qhZGV2YmEp?_NdwmjToE) zE4qG9$IFFxgNI=DPtrdegVQ+Q9^u#iqFX6R9>fp#LV<6ecMPOi<>)hez8+uv4JsfB z_9>SO{0H#bGr!`0S-Ir7%!W&6LC2kO58`f%?=`WrHHu2mD!YB$BR0?QTfOVg*?5z_S6loYOw@kn;IOWMt^_NLk#v5);WC|nyv&C z!4EywD878^adL@7n|%K1|n5xNCeH?K?Dp z+9R6duNXRWB$#MxczLK_e-ZvWYM)Gho7ER%sDW+oI{Tr8193IMTkqm^*X7+#jL~bll6AhOR!tv@$p z6p8TQ9pei;oUidQvW=}&(ja#lzjTOu%wp^e>{p`U0qLRnq|2PGF0I< zSuUt>=sU#@ze?00PL7jY5?0izU2o44+NH7y6tZcbvpvE)wMEaLdvMBDJ)mY9xbpi~P@gvycHi%VX%TTGr}iscgv{;Mr9N;2 z>Og|3ygv*0Ov_xzi}eOwPl$$xBPB=rbW2#JLsY(}P5AcpBz-=oM`gx6~wVuX!kA1Q2?7?yw<}nLoYE$9B@zOd9BNV6&jZC`MT1fX$pkHx$aN5P5)dX3tMo_6=YMW0yKzB($ATed8B-J0F`j zaJgYr$7_Y6T-_v}VtAq3(;66A@Zp{ts9b81`<^3jT@qGr1hVo$@`L~a0=aqq#w0oK z8Wp)b(Muzvf+K{X5x@fFP0bwq6aZxOYZ)ebc+PDdZ;Dn{J=0z;qr=m$*T;u0(7!8w zlgg{((R=o)62CiR>384CVuRz*jM(<_Km=>AJ7QE$Y8HHbkP57OIAI7hYVW~k9>sm_ z+%s&RvJK~PfOFX^x(#0)AYS&VMB5=jBUW5kzkxZ8M8G-u?0(Lgj$gQz1SeVVYdBy~M9J zhz;`dS-ow?w;ZVhBjs3fiuY|e>g_R!WEE#cKHNc@xLmMLW?TZ>9qc4B30QB8Ubw#b zsPR;XeqY(X1)$3*svnn*fqLxjUiTHpIA7rYX|Hcy8R6zc%v`qT6u{v$N&`20_E%bZ zc6q2Uc5wzj%F${?uQzEdObi`a^ZV$Lh)CNUYAlXP#Hg1yFlfIGvsDh3pbl z0MVfX>7ZeL=37P=DQa@aqR1D6*`Dl-8;f^ke$T}2pP~Jo{~if!sX{-G+D{C!fWkfi zfD7!%@Al{lwXKkrB}p92Og`_iBXfHBQe}_tZL|R`V&% zUp9nN5cIg_2FiuxfC|cr>+++>$GwcW1DjgBWlmtG;upZdJsIl-``&wiHxjSI0oTH` zAc&j|!bT7_=}6R8y%%1XP@4~&#W4^e5=)Cul@t$WM}tCYmIl87cOBW1!aadWDtz4Z zc)RhQ+8%NCo7!WDWDw=2VvcLT9xvGFeZs<&9I@E-Q|ogAW(D9A9^s*S9h{@}K~mSA z9pNHI`&`XQH-D(|@i$n&7zvQa?N+imY7UZ|5M(To+8=6)R-sG?4bU{8w*{jpr_W!1 zz{j~ucN5?wA>x>`9D0A^<$!ve8PW=+!-P`UX=k6v8H`$xQu%T>RyVyI3@sGicc*Jo zDcz<1Jc|i6Ux?{(kUqD^{wzsSQITB=d_yrS#av`SL3l9ZNyjbD=ULpRQJz$IJOTkT; z+&CBBsLh+weJo44K9kepb~nGi`CXm~UVstWx@7S6*^fqi6E?Vp(LuTE*bSD|Qr^<* zNMJE%aMpO0y*(j3u@Fq^pkmJImEAu}2J@8C!qoQy^)JRypDIOs#{()tdo?BV1SQ^i zMH40S?Bm#7m^)%Vo?_QDSAYlujspLL#AY|XIZ-dr_Y5w$XLja6DQ7hz=9&H|ywBZ>F-f6iM-6{HTnvI0nHxC=WIw4vP$aZ;(!~>scDd9q*L&j$@N31dyxt6bya$0LaIe#KD1WEtNr^-dui4ZH zhv1H8kO@BkEa>ms9UW#z-{k@aO^#d710ID`zYk$>fqvBd8a_qmv%MDBplDu%J&gVN zb{NmZ5UE$keb1? zgtE*k0{&etEfsB_9+h1umI!R75{G95gdC-B3}-xBrJzyg4QP*J1=UcFKj9CP7kput ztAEF&{=mi!24AlB>`P$cd{R+ut{`6~Sl9b<@|LuNl>sPR4rsZ@=teG%hq3WjHhQ{; z62CyB;C^DwD+WPaQHNuG6(rKrw zB_@I^yN8Z?@ANO|0wOtb-_sHbA(Jay~3d#OK==b5wPf}s#6xAb&74M5Me>;7ZA z)kOR^KfNQP;pC85dy@j zy8HbzguSt&ej9PVKN54%1sY#yRN6l;*=!_#)E{PE9eYVW-h|{R84`t@S_l z?Gtj-t~uIUem$SYJ!X0+AL>WAt|Obdb=%AA)9b|+XNv(T)tjc}QSo1_d~z{Lrbk#< z5uB_JTj$B7a9nBcWour)*%9d?ewV!(((|T^0iM`U9}L|V$0P3*7faP6k!$Y^A^DYXwBb-;Be2&iWuk$C(}s}tGmZQs>ci~o@a6w7vJw36C1uGHjlj) z?j6Jg!osN%WlRA!E7^l=u3nAUzAl1F|&gN_M5BD$@~DVHOnx?eO)Q(;-)n=t?Vu|2Jf4qR}A~YYDb#%vBZ`3 z7NCci?#SN$oRsvLf1#Liht)kma-bjf=n6asl&I9^i1@6Uy=aK0>mGW-il!NyLA9pz zp8XSx-sS#i;gs#MuMsK$Cbd7^shp)}Gxzm?IG`mV=$xZ}GJyz3qM!ceZ`bFIft5#S zF75s(FoEy~vbU(k*I6oZV_&H7xB_!r4I+*WRL}eL0=22Uyf)iS@>HdKN$y^O=B3oL z9n}EN8VzQ;(f8`FzjvRH;1=A6Oe|UvyMCe^G!P*jPbUsjakx7ne2#C%{ic@t!{}vX z{dI3u$!z=xbF7u)f@u10GoBGHJN+U`t>bu6zn=|0H33(0iH8b~mxb8OS3ykS_9Exy zB^gnb9uaiP$eVtE&Oy-v;ZYafPj3wCjF(uMkGyhm@5{q@HZ9s8g~M7-6m!94DcYYH zjQcA+Ro*CEIkB(n>p5$I2Bi>>)*|8+0r+C=&MvRvJ$WLPjP8CK@yFo&N;B#tK?C{p zTB!3C9l3Xaxqu6R(9iGm6@SZ_uKOV9*FhG97gTrqGEx-5LVUhLIiT5x;bIuyM6~A( zbH}CXT-ofkb6y_(^W&Z^4$-lNQ&3R4?@_BRq!U&Gtmh3Py<*Y4^uTAnCe+qyC;sV0 zjtOkM!_1#hrhuyBa`rr7-w^BTvk_h}*}h4F=hjf#14e)eu{9YWVhkRMyq3pr1%g&G zAC=bP2mi;Ou-394;D?WhH+U%4S9k&((&wM+qSiIHQ2$}Fch0A#z1oxgIa|0hg5w<( zcpQej7i;FGU#67d{j7-#v2SRET|nGFZm+9<*~j=_d9IsFptt(k3S26I*snnyNTLp#?_B} z--&92II9nAdwrEJ2OLw(6i)iV-{hMQ=k@IZXL*;}D+cWK5zp0XMg83}uaa)QsXdoN zs_v74OyWzHy|OFps-KUg;rfHtFWD09 z>-K)z%$L|(F%MazKto%<=|(xzRE~)Z{r{0bImk8ImWJddlCMwna0=MAT9Edt)@sym z=5%IaFrPt6cc3^pg$ecT$~l+y8VfzQ!qNpOt_xw9PW2NThl8^q3lO{2ycWbmKLDJ( z6ta<@u*gXX0s*9*HMnhuvI$^(*6 z`#=dxyM-XmG5kf{UcbjiF8gib@>$F~3Cj7(=Pz?gs0B7WV|Q^{{rMJehr>l0AX_3} zCrbJ^P<_+}x&g}7b?3Z*`LH}rhsRI;ib8h*pVph88nRXu8kc z*kd<}n3jBXrmb1Wy@L0SzQ>S0b|=Y)6>C`r;v~coLpxh_+pi{^Y)OsBp3p)Zpx(OM z;rM$5g&B=j1Z5^6nTWUd+X1uzC1|wIltbtTQ@r3hG1TI$5>5n655~S8e|(G4M&+^m zwnIR^A%RQhauvinJw^7ckgx*_GUt|^qeS|hwL)K<8^veHVzH|BDC_kKBtOOOR0Hd@ z=WRUh3%=<(Rkb6+3)IQTi>L0{){25SvX>NIh0dG}cL|ZpC0OCx`IE0W4>IVpeyE1I z=>E2a)#vS*^8_z4#P!R>DL_mpM7xlE47iXy`;4@xA=|xRhpDppuiTCx1mknP?D^-8 zLU%O3MIi711kYTt8}+(Y|sH>Re`zYXEs-frzcbuhg}`V(-5Z}M~qRT zRuwDxMON7JX!Yowb5lfJ1@lLG?7k5p7g!(4aR6&cJDuFHh-eH6CsRrqe+Fvy)G8?I z<4Dh3R2gz zsD*p-B$LHC(U2!p68NR7KknQ4i(0OdM?O5vsfdDm$bChih#fSEkNG*HapbHY(`+xe z--PRHK+W3_0)uBTh?}EvNqeQ^8RWg$E#Ao_&Tw|Mgm1 zd*)sB*#|JMnT0R!Ra+TU$f+){hB-~EI+RDYJRe$<6{}+sw4UnaZ(8vEWjF=IkmjAy zdUT69dDZm0DIwi2Ty8~3VLZaQyx6t5010~a_oZe3G{?&dRtnA-H+}y2lk=xxzNyN8 z@8ug4?+!}uqAtr#aPn^0KIGFWIK`8)8+8F3o~N`5s52sVO!one9*;Jh1IKO)WbXMU&ev^E@dOD!?|8>Q zTKJhGCa&-qOmof&?X-_x${!WON8S0_AB~@t924h~U`5L>o)grR?i1KgKx)OH-Rqp* zp)-1;&_V3U#tqvb_j>!danAu>U!m7qvHT9bge>2@4(v5E-J0a;{P`py4j+@AF9^?B zZ}QEqA*GKn#Dt57SAoskd@G94tUUr9eac?&pvqcB2x7?0bbs0x;Ecoi$~r&H`CHG> zIJ5SQQc24NaQMoM?D{%R1)GPG>Cnoz$zHU~#BWY7pC|OopVN0Sohkb5XRc;|!9r|@ zPc^An-Ua=0DKf~{-=*3=#$6$!oK!06StfZQn;_NBG?=GW}(7h?a? zVIurmVXZ%9X^MeC%#oa}>B~UW?=7~+iVmO{99LmxK-e!|_Vv5AEn1zYs7(TB8jmge zX=$~eR+}#R&Z*2)+X+&CN(l7}!{QCpUC*y3I5f$8!5&lgA$k1*&o z;&Z6RmaSlt{Zu@$l*a<*N7Wg0_5B_B+c_R}2V=Q+C!HbRz_X^)^9q2e2sJxe4LxPx zYVaR?7kQCLOVYOCiWXX+t2+)F3Ngb0!*%xFdN4s z&+j{grLYegE(v?`z3KGsqT!hDciw0rNrktWIC#{pr3!NCgrec%{Ii~~ID7m7pK%_zCQHr{fb|$fqEjy-p%v3iqD-qBz9!)!c7> zAH<@7Ce4}?n@{rk<$zQ)eD}cjtsXV=hw8oeTf=I0wr~6_G9=>~O%=0+Q>R-u4PHa|FdM7L$JE z&G|c#igJu6%$!m989GG2*r@|ul5$LYq%e54_PaP!+SP2h@IP2Ovn@rjD2V};`RHQ^RJVG~v%LuIdfONkW$O>Bj;z4cWtC#him2K5z5G?)oa7(ti?qDi_%X%gt7*shdc}5Wv{6FYrgW z(N2$}A*RonL&9wUCmuYG(Rzqt?0q&C+=jTuFA{WH4jl`+ibD;=!wV>yr+?|86W~9; z6WR-ijWc;$4YJ7!+c2IV#qo?6c)?9@lnltq-=*%LB0q%oaUC8xaYh=A)XALCiq3g- zhG(cx(1^?OVP@>@0cC68-Alxf6C%@>0$cmIk2&sGl3hH`Ds;=DX7Y;Qtr4obY%F$% z%ZL`Y@N91xdNXQ%>{s0NQoJ@Fs+hn{SxorRzY4)MmT_> zpGe@y;7n*Ovo*Ubd#>;O?wSO}fOJVf+O6uj`zrg}H>Rpu3dIE@1v{cc zBBAJ2calEgARGr0Fn2Q$w>*wA^suHC7>NG{4mRZ5XQ zM^lehJemFjX5lflQV&#*G$N4AF_MCKP7<-cB9*YnP!Bm|T3sHONaj*qK_a$+s%&wO zyFOqOX&Y|P+H~r)NGBeUj+3|q0${n!fdp{;Pa}gf9lE$K`!jiuM3kNMqzpkCcyr0h zs*qUk#sudF#DuOx2>Q|7oWA`nh^V-9mwN-^yx0K$)RUB=-WXvWieExJ(AZncb}%H( zac|)}WK7lqE?7kDVyynO@p}0P$$%P{;P@#2hw6a3rsq2RhZlO}D&69m%beOQ^;6TF zKgk6XE0hIk>m5*OaQ~1{3fPOddZ-z529RsL*XoE=KC)eh44;B}T@YQ{OyJW38M__# z_2ze9F-Xq9NWJX!Jox69aj+vtonipkFq^xp4)tqy;cos5r{)8?PmyIpvxZXvR#x6T zw7fsytQvPw$i=L+xgV4WFSijOgDpP0dYo!WmBU_~zSgr#$JCWW-KSdUf?inAaaUJ$ zU>&{pDDw$;N33nwmlNQfX4wAd^G?hT)uq&ThvN&IFfJq$iF)siSXmQ$B-zMu!YcVb zU*W%H&Uc{$C)?7hCV;8EhB2v{0xb?+uJ4t{+(*3m`r+-f5qm8=*Q?}r3C)@(>!?^E zhk!*vs52nj{QbR3iruW_bi|TlXyW(A|4Mv=quld{Q-B|XofiLWtr6d;U#b?Fq1D{F zzyT{HgSx|!kLi@(g9Rg)3!6dJ+QG&rB)smdNB#ysZ;4z?Vw2}Nh){JZF7ESw>lk?& z$@&+6gQcsPujAox+-LArO!?>OS~IzzeRLL`pWbuvb;g246HBM`MYgpEA~^W?G5?!? zSnZ_zU|*-ZNyHe3<{Mr&?lFA_f1mkLnYFKH17Rq)%75{*;?mXr%P@crS+FlL=Ob>W zb55_*ewiyHx$Q~i*pEox^K&~rmF?H+>=7lA?tmn)1s&ND@5##?bW!lYeTuyfd9`n% zg&dY{T|12G?TD{hz{J z&BIHIDN`bVEy54XMq9N9sb8ey&saElnVRO zYMCLW1%CGl*C%@1M%HdsuuN`Dlc6;ES`hH6*Jz`pYMi9!AnB<{xfV^W5-zS_i3lju z;YG>eHYtowceb;8fp%1{Dy`!@_xbY)OojmmCtK0q(e*6Tm!h}5F%5(cFrFbiM*3`u zfh>t%IFv*iQrMzrSxTrMx0a51S&tdd#WCSGb7!g^u&B=&`jki1d>^-;yj(!xp%1u^ z=GziGG&CT=5M(m@1sAHww|&nG1r6VZPO#hpAhGwEO7~YZK7wA-Kuh`sm&23`a0;QX zrlAE(8evZ@`-dnS(MNJ0`#A|wRqAiAk9X=U_t)n;lO?|mspO73p~S0fKq=(B@%Sid zZG7LTT^w+|-x79cI6b5L9-Obk1KS+Uyu2(fxE)Er_mtxp1_lr3-r4o{&ZhIeD;Z8M zFYSChVTOElQPGs@ypMY{^{(5pz?2{_tM0%p2@=KYGJeQUj)+2!I$ZR>{TxEA=%ksG zKrm_4Me+9Yn?T()J7PxZm>u9nK`r8f^0$cg{FWU%hTNwy37)K`1jZOuzV-Sp&rbEq zenQg=k3SbqEjk(sK#h5sT#iP~0}(i_-5q9`4hRHnrq|L8Z0Fh8@^5l)59mAZa&Vu2 z!Lf)wvxQecsa40EA)y}_fP~Q1QZl`3uxzgd4lX9XBQ%$N!ir1BHP=9neb)=^-9^Z) zgwRxr-9{Qt)e^ryxMjgCDhXG$lK}~>)$s88BK>+(FbU}tvR|VOwIUoO8a6c zo$`uT5n#3=B2Wr1Kac7U{H>w?jEk#)Q*^?@sPDlwZ(>fuq{Mue;KStm5nFwabP;bL zclI8#05uyi#Mfam69&16(47Ql`zm*RRX`1v>@%-}M)eugWYxpetLuapR>ge^4d@H)TT29=hTs%o^O-Q+MU^uTgf&Yc{#yg{*5d>~HMsOrQ zUo+%$WvSe;YsS3p;H>h<8}wge{JQ%)U>xdq|MKhcsSNwM3tzi&cDN*oMjfguK2xlo zkx2aM3e^YWvd_)*#RM@TRnHiUKy5BmTNeCf4}^nOX|AGu!8Cm@hJ^m znFQRfK1_UY17~oPgCZ5@`fc2Zzl~ucL4nR4raSnSPXe8*Yp|Eiz`$41fF>BJ-uO}L z?8jxl{xy;?I;slMh&p}{0Eizw}H_o+-X!!JCP<3OLm`jm6-j zOeM|v@`VJxah((g-~g=192V`o>^AW92hM4o$N=XZR{94IEtdKH;(a9ipbE@qUqE55 zEO78U5RQH0p=&uw)%G}Cxk>HCwo@~jAJcvG&z%+)1N%iuN`h)Fx{%ac?y(HF-_`Pd z*6EE&jYDJpdqrC{QkJ%K1jK=jC^~r85k*GiZkF1D$x!d0V>J6Ymhv6qlEn4X)VQCfPltUAHt!WUsSfwe$||UY zDNz=D5xt_Ju})6TfYgC}+4AUuq~0L8QMK3~lG3C_LtJ=tOQnr8qp6`|Ox zNY0EcGPVC44j<0TOeU18uN77tDk0kb0TvVbC<+K#ARi~n7kyrM<-j_0ku(R{GSs10 zpC7*!PgjGIJ#+*6KLB843Ma?edRdmTuMeI0o|y-Pfhl{0lX7?{{K|331^1;pX1wksp16(Kdyo3+RD{5;h|{kk(Rz^Vzy9Y!4c2GF9Y264v; zX;8}r78^rNW0)X^Yr>JbCYrc;W2k&Ty3k@>LH?M0-7@wQ=xTG%TWsDms7UvHAB{7( zEuCeETlm1~C)pAz_6cciud}jbv5M6_A=se%G*U8g;^E1bRlcfNb z8e^C^%6Oor%$B>^GJMG??U-Z{e_LiVH6=)Jput+}RRCy?kUomgcHIn%mQV6YQ4~xE zT*%0bcSD%LXDj`9>t(FSOZ5-!~q$k&oSE~%TAwO=9a0y!Q+Q-iUxGd9hduMc^CUFX&_BCWpdE6 zFZ>Yi$bJs)q6)}upDop3mL(&Mn=r_pIRD_ATlE5${MV!!(2`jw+n+0Z{Lmyjt!Aku zKq;;;oUt}}V(}$KADqow?uWlD;!wgGG?5W}s8Z*Q9?WS{lMqKBy1K(aeU5pFZm2d^O z#a$esP`zqbpNj!ajhz_dEsgrG!)~OnyfHHuPiv8jno;F zo>T~>LS9#<_*k4IVUAcV1wn3{Yhn9lNN30ME+qbE5V!LvVvC)8a`M{d^|HRU7oF9v zx^Gj^(o0RL*&L-Dv!5>Ur|3$qIF}-)^!853J<$We0IFzg%g|l3<0J@(XL&XfEH=Ls0r<-$t6-%V6>bkUabYdwPk!YqTPz z{a&t<(lZ$Ox1=VX8x0?4mma0+A@1ID0(`>46@+SlaSkw0IPD=d%t(BQw_bj}=~Mag zr<+?u7+aKde$wKwWIa&TZPFl+0PbE0i=ON4?Sd;*amu~rzMccAC)$hIXnugMVAYB7 z-R<$N^qT`VA3Vq%U$?-N`Mcv#ijUc1GcFw;G+sKX3}Ki(3n zR4%$bVfm#;I;Tz~0^O$`)_%*{fu2%bC6cJ4JJNRG8tE1qg%YxiHc*-O>ckyRW)5og zzXl>+C40v%74V*n5F4DD2RL)96>VGty9j)Hxx^{7*Px>u9lcsgIHi|qENu161u0O8D7)|;U7|(*i z8esd(Q4x7M5sj06+u0`ND{_9^LZcc}{RA^iZCiD}uJhwR(MYT3%Mrw1Jq_k2eo=$; zD$gBAh4vu`H&QSs$u%?%;4^s%F9#x}&bp^aqQi_H-BR;W;Ht8^$lvx#TRFn1iZSl8 zt5`9`S5EmNcaOkvFo*Towxj6c)NA_0<&0nb4&!6+{2t(4mlJSoz!?_c)dWmZRdLSy ziRUrA_Fz{K2=SeI0+!c(g-x8?|FUq(kn7_o+f859AN~*m19eMh2bP;y`~^Mg51%Sr z!$Yu7UssxG@To2RXMIoKGwh`?R9R`&*B*42z&eFbNYmbUBxpVWs2G4oD$dwO#_c^M z%VD|6waFH)3Yh@9=({7ko~h*vMa1NBz_gs>6a-dfA<$<+>cN2biMYu_b-D+z&my17 z|s_(X^0B{P1ayUf_n2h28i35~vp@fosVVQ!z z@tFz;5#0(77xn1W`r>oxA(&lm{l2Rim{~5}YU=}2_aIPkhT~zr33-A|{C4l(L5LF0 z4bzN%#ar0#!vzua8g3?gfb9j8&wb=R;qMC?2EY0DT34j}neOM$G9hu4CzwqXKm^`1 zSmBiVlyTAy*ZamogxGdCyK+zu175yDBzwHc8k1RvWR+ULsrT+J*ayYp1pQK9U|v90 zEHE8Da;W!H9R7=_lF-!wm#pf>0>Rc#$S~4KEDSjlH%^fF(Y`*fLUO+I;v4c#?iUPk z`3{Cto)J?pL}TykrF%Q5Y-&4Ky*OWL5XcdtE>(^`WbZt#{dL1^FPjdm_>c4h^GaUx zzG-q9PFnk>`ELAaETv5AZIIc+XTu7Ubaj#z0oX!$1nr=Lh`Fv$yG;(@My^fu(>oC! zXZpl=GHRL{sET-|&XA!D$W`&eII6{C(+qb?jFd(%`Y!MLX+EI&84kP?1$>8@OgW>p zhT_23n2#I*Fu}%$iGm|uq5ed6!Xro=YjWOpyie!}&_ElP+)xC)<#ZT^t?`+I#T171 zjN}zqX7CpFWe#uBMtj(Ps|S5P?~J+q)60|rBK3;Ori)B50;CwuTTOfQTi=Yly+B*Q zOjHgom1gC)$1$EX$^kcDM|Q2FgX|7@>xS?K%QHS740p~(6|umUDn9nMqd- z1tkRp|4C6VVnd^d_Z@5R_hx}COiQrzu<#uiY+LZ6bxtF<*Yh{qt|wxfkELGml9n*% zll0E=ce<=e(1{Qd@&T>NQTbT_+(H?lA8V`*2J=mu7d-dta{`1#g6?%{O{Nq< zLD-<(L-C}}F^_P88E&(a)l)i&8mn-M?zo0>kK;#J$;vY;W$c06{c>HggJ1m$_|CF5 ziI@8o#{svW^yY8cwZ9hR!D^U;VzDBQVeoF)QVQ5=a+A)0Hc#JBFuVCNANL|9CcMLQ zWdTt_EaAtiphSUNd+xb0?JkE8SoP0deSuTMQ8W2@u0L2S8LH9iy`yFWrL`ue%Bcp2 zNDU#x*Ye#E$oVD+4P$%Bw%8fX+2sWzE@0XaH6tz=sL(lQOJ zjNVi-M+N2e`9<3Zl1^5`)_n%A9<9O&Y#6u*!lpQ~{4hpmqF{+D#$?5#PC;C+l~W$|6CJ(q*j>pDZ2zP;ophwxV>Ow z^)>W;QDp?(S*dawB>R#Dn=Zfp;)Kxh#*Y1cWN<${&IJZ>pqhCz;cY2y+Iv6kEw>fh zBarn)%_*JAPyta1?_Wzla-$7c#6EAtH|R?Fm6L zLbD(Tkhpf=-B}REiDU(FikfI6l4GFb6t7ks>~$Ckkf0n7V!BMVSR}=RuGz{TnPQ1c zibLW(-0UZhe*Wo*+*>z&A*16d6kQ%@d`{JB-U8PjXs~dfzYjm5PP(QoH-isPP+EZUK%s>A2o=wQoF}5+DQ~L7c z)BE66CuCgh&-x=%A`jvnno~>24-1&RRv*tNxGdy%qwFFx_YT6`jScvQw71RfY^l(y zY3*g@a0W(9I2gk1ri6w?eb{S$6S-FMP8#$b$`{8_#JV4}`;)8fmmV47YY%Ybt~e>R zt>JxtjmOg3V;e5#r(Biz)>ht|0osNAXjF?+1eW!T(a62Z3I6G7A095S_Q5dt(Pv9D z?(mT8xP*V>^jq> zh+(K?NaKDSj1>Mp^bWj&%BLlq>EH(@FkS?c6`S_2^1hyLhgkbB(M|RFEvF3xzWMdp$7kB~l;qQ+6NT54r;GgIAN_t^SaQRy3c- z@aX6_id(M8%lDELd&IGX6@7Sg*VU*MUvFKc$=G1f+w6{ZIG(cTa@Ll6izKb zev!v#Gw@~0V`cLNQwbgVe}quuYDOFnS&?c;hKk8#q3$DJLAJWl85Qq zwgHDkjUUV8mns3?j9bVDm+&ImRA{k18dln}Lm9swtNlkLZVyRpG-G_NQ`geOe$__S;Ib&H{=B2Axfi|u z1)VEz9OgmEEh&rC_BrA5w_C@@0|*EaSIp&PAp47L$#ReJR1epM{X@}T5z1eR=#${dCIDR(#84T?WPCB{0vL|JB)~l##`{woSnt!PxO~I! z-U_<4JR*=EkOu+C6`*ne7F>``!<*S4}OS9zrj>c>BbQ1#zDR=1&JCf{wrdLkvGh zbP!*aeihuqJ`rwUfxN{QW`PS#Sz?W9upuKTRD=$Sh*PYHUz`jyZg865OXUai3Pw1M zM5BEvftomLCl}a1h5fegczVdkp)RSA9dLc+#}|YN--Il1Okt!r(g8anT>Y;Kh>M|wVA+t9y? zs%i23Ejl(RIam>e*PC^K8oUFG5)^P3A-HTb6dnk2+yrEZ0$420&0kCWMnLIsi{Z#? z-P3dJ?mko*=w7yRv|&?>!65nE_g13^e)l4eUn8LTl`HkG*|UHxblWFtdhkXWH>hOF zKcrn`?ehk611$o(fb6m!z1Y17fZ=l>&&j@RR2lavQ*!T@i<84|-5|lm?#%Z44b47# z?N3Fo;Jv3SdPO7aMD~3>kPc035KL!7q!y)}7^R^*kE8sm4ro4!0c_{@L@z~`?L8=B zYwG6JarLz>y&Q-OF;$(W3rbl!U@EXVP^-j|mg~|Jx8Tx+r~n}A4|VGO)6Ug%%cl1L zK!?TGQD-=Q07dlBxnF67K1?I|!kNH6EdC7$8G6nv&kZ!&AbjVZXl&`Mh?Z3h(WGt^ zO>QcT4{!#3e(B2?F9=Ex&t`Fs&_mA&rq)h@{G&e0Ey6473i@jh$IXiKLVFUVWaD;0 zFp$JdY3jouaXV<|IUPm=6@vx=L0)2GBI(m_P5HfiB+hfZm6?{tS+YkZ^dBFtTU^gh zdt^gQPI9)cBK_Vn!-Tcgl_KXPayNtV3|NB&?m!U1Uc&k)FNy8BqEXY)?54>3a6l7(YS!nu&6LX462m=Oo=c)h=}58V}3M+{y>;jC_al$i5;2P-H*eqabOJN;3tot>yIhGTuKBGy z_%go{mxItQAn=Q&Imn_~;$v*_G1zNZdI#d`bW_vzG$p!%Ur zu07;u*I=P4-yO>kMKvSM5y!Ld5Ik{af1D!42m%(>kor+Yx=I3t9R&>cN3jM@9$_2x zIiSC#NWh;aIhBi|&-XKYjguXv=_I(4A_Qw6*ovTO-)5f~&g_Rh^pB8a)s!aEW8kCn>Z4sF{MA#{=_J}@ zgtH2m8{6zH$cKo_DU! zr4R>^4#tNCZ6rVRcBut}UCaVPT8D~urBdbRZ_rrK)UgU0yU&#UZ8AjVU(xs?Eec!) zT(k`+9f5Q0_Q}=EE(jFP-X`r<4XNnU)FOxv5t)IHNG@v|CDTR2Y zTiP2V1m*Ar@9swlM8x4o$0q=Nc01+vZIjMde}N|00|*lOu@n*+6M;TLs-T^730=8z zrcnqFKG?%A^df%OjXPhzw@pArM;dIXD)jIkr+4*qgxGnXE1vkqmAI}i`e$(0x}l_f z1HAGKwjtJ;O#5Sl3=N=0xI#wUTsU^6mejTTrU_-U0DF9`zfHJUwnMb~>mAhwIA?T;oW1m@pJnsJ$?=5QDiT}e<6nY zz5yC?nY!u+D+Hzlh1ZAS$H55ska(nggO^d=LRjx&TD4a7$%R z;L_GABX<4dqW7qaz)6dW3zwpGNgIPy#Z{X)^s|I+#!B+rZ=$eCGFFu38&uZX=lA*q zi;I1T)%84@BRz(1M6fdemu34Jr*S%-+2XHHjeS~?wd6IiJ&FA@Zi5fTHC7Kju6RF3 z^3C;`W5D)smGe-Og+-EWH}YLwu#G+ZT+W#u?eU;l+2eZA;}KESjktsQnj!GmU6W+Ppr|%`Z*}4@uu#(yPBsAHzT7K2;b8EHBhG?ktJoXM}Zx1+TpYFbRoYG zZGIr?RQiByk)h(f^x6@ZeBV%0S5#(4@%&6K`)iiO0V0m<>F&FMp2Nm{@FQ^$FR<=VzTCATvpUYPJb%UgRz+U-qsd&- z6f_+2i^bTyq9mq^@garJ1M|$j-SfhO>kCy-jt^8_1R$OtpXqYGe{ZTcIT!}tr3oHl zrs$t5e-P5*vm)jBVY<|foZUHdomE;b$CxCV*jo3c)rRP>tH} zSbD$R$olhOH#WQ@y=@*|tb(xFPf=PMKeJ5OsoK39VfRU_>MfMrJvg|wyL9P?&Q8{+ zoSy1kxY%FGQ)8ruhKUJDD(((@R3FutfaMlJ8}~agfvp7{G$ajjX0NbxHy4y!O`k=; zql!(*;A?_B3<$k>7w`Cagv^>314u0vD1fXZWb=X`AVlG^_oX3UBNgqkgYzUB_XDTS zCXz_=H32a$I5kjkVW<8#bUkwFej0+_jB}(?`G32@v_#sHp z!jG%p-)lajwzBDvrvwxEN-eDRq-jQMd57R^R?UhcESq@;E<4B=|e$^ zgi~KWhpgY${y2)e@UL{&(;#g(ri}W-w#xKqlp~(-7$*ony8TTwFX8ZiNBDqEX z;YIQ{jgQ<~Gr4zL*?lFevEx;_IO~Vu#$oR0_jBzTAV4`lcG}24SUY>ltAjP*uY|}1 z^C23a&+$|OQCFJ7SE%cXIs}>*)OZ+cTIgxA!)4y@WI0kk=Ra3HeWp8;5MT*iE$z!wf166{=^)~H+DI})3}>??LJO2 zVB)W9mE25-K_kE zI!-Zn0AZ8_fiZn5?Qxg%J8qI<+9dh38Gm=X`MY|184u-)pfO<$ZP;eo2;%`#{v?tR zJ{YIw_udcL{P)+^tuV*|-BjqE>fh1(OLbiZKj?Awewd&ir8WfR3KFgZO%+Vkz^~e0 z(DnfP_M8p2z#$vhEiu!dh>U;)ou#I{L&qi?budA>a31v^2t}bo^Lsu>Tp^$GEy#@g z+4AQ|qNE&3A2-PFWFL3!R1_;!@n-zKQjG)wXv8G@ZasY#$2W(+u#z7a1vHUdDjDY8>FKBfy}bYh(DkHarZik!#adqqQOf?iTn-_vfg-?5b)7! z+0-JL#UOT`v~{mEKdzbLH6PN3m`+z{gXXuG+L1S_iZ1O;@8R3DA?v{b;dc_H{yxE| zVij!YW~~uPDT!4XA*i&#^0{i0CqKZt3>5c!UEc~AJDZBm)=gyX_(`z0=Ml#q%D7j* zBTfHW18_myUsCa9a;?NkNeYf!4(W+6UaZSpLcr{Dzt&|4j*Smmjt)F2E~^|C8$D%a z0mq#}c|kP@h~0i6yQ|)df91lfS3!G@#~(cy!tr%h{6y;*aUoQle}nj%J5U6k zpKsx8d)weezu(g2;>~{0bCbocE#iD%X{I|-G26IY^m%!d_2ygeN-6ugc4iM#hEQa} zJA}Fy@xkP`cM5rN>F|y&ANMF&-&D8sv5>4VDMXZ4uT$_%w-r5zr^kF;7%h2L7{2mh zxOn7V5m3(WZCa|+ObF&Ba>sGe34@ZzXo9Lv{v?vfQpX~i^LSqdmJ% ztDYl^4MMacfU^Z6_qj`_qzfidC6w-(LxY^Me2~_9p6)z)M9>|>ob`(KJS-U+H`ZJ8U@)2Q#{DjPoS!X~$z@8I@8qIG5aec=V!DY zx=cS$kK1woXO4~ATZ}+dH}Egk{^4QxO9lgA@EaUlP%RnWKWtPNBoBLO=5{f`-yIxp zO8t(-S`voSXi;0lepy3LC;%!HWf2K7flRMOBsg(-L0_uXI>Mv+z`H1TAr4NSVbtU4 zCk3G%th(Fvf1d7IU^LZvf5@9^gM%2pr(Xthxc>^iRd(KQOuRYRa3l8f#Zu0#3^etk z7_u^RiS}=uS+l}<&{zrGdD&Q1B7`{r#}r+E29kcUQ90RKe+PUae&xfXzuusidP!R3 zsk|B|YKeI07*8x+gws_g`I8BqQo!pt1+QuOF|q?=mHB{7n8NI%b1!Bs_x? zsG*MzQT5x$eH~sZZ20wXD)}Bp@6xCwa5nsXPqjfCspLqaY0DLS{U{Y*z4f(jyUz63 z8rTVWr~nFTFopKvQxl)cQ_)c>2PjWTnYrD8Rj`~Fl4=Nd^kNo}fnt*i2H`urS^alMNjWc=C^{nbwb*nZSe?I@z* zCF3_4fX!S%Q+a-o!}qFdWTl;vHs2Ht1TZV5X-}{wZcW!TC57tbAG#XVj?A(+|L_eq z(K;Z>P)is06daauMSh3v6eSI{u=J?q7Bx7B{y6gWR3zjvRO(4rBnwn~0`n=DtbXG2 zMh_20ityjh1A0<&lI|fqPCtC)+*9{k)eeMkZhv1jYyHANUMMbfnBapuLhFP>Q*-<7_<9aOE8_h-Rq4Kt z<{Wg&X2#C)M@6aClTa4Gk1YicS?5ho0Hkx@5W9QFJ=`z#&{F8(D^zvpYd*_tSyw?! zGjp_?w*h0dhZHBd&YTNSoYQGI`8J1 zJi%o9QIy0}3>8t-e9sOBaG3cK>C8YWt-FhvJcoYVW|Bu_HX8*FqKJz}bAXMD7UaEH z%U$t5DByDmh?>W%*+BdUc^WDv@iPAG7SxTPJcZ=DK6P&$mm!AZibA1KcoR~% zcbzaZ`y=zdi4Cb-zO?%|WWgCnQmjVsUwLIN&oIdJ0|!KIFfoVUIJ&?1cD*Shm>uP# zc&!)S=jd<+G^wHz_TG++bWdLqKHl|q8buurRlqn6x^oS%37(nE<%4oL@Uh4WI5U-#i-O`%U`^~}?2GXne2MIDU|_l6+J zjko)8hr1*2cR!@;^3YY_7ydrnlGNnyHb~c{hPc9*aKVey{tGTtcGr5B*B^oOY zry{V~L~R8BzO7Bu=vb=X?^g2N3VzR2lG=Xy$gm1KWDheYtmDqJMk3&LDaUdOIY&*g zsPDb0g|WwC=ZQ+?t)LX}d(TdT(l|6fQLiJpqX7I&u9M{$zUutq<;Qe~d?Dj#xl~N1 za!oWO{E}Q%}iHBMV+O8((6fS%Ly-n|MqpBXYu(-Q_rGxTY%apEJ1S7 z#e@P{0|~mX*S)w3DAh-05~@MELh5<~Ab1!__o%C1x;k$az{objg-`9~&g}@Qa^c=f z9td*XldZiZ5|PIyDZ9f`o8RlEsJ{6T3crdQ-jp+|`@v-Z1L!q8#+HZE$!{MSt8j!! z)0N;hjIIrRZ9hJL_u&dJ)cCO6dnQYx+UUv+1GoWWqbjJaNLPu!N85d5J|P}7GQ_l8tVIKU06cty3Ia^(Dnte#s~@fR`Ed`k?R zOVqo+&vC!3D9($f%>o-7LEgFP!uw7g;iw4OlWFOqeFPj8XgsC-YhRsN+dMZV5qWD| zK~&ahs^sbWO*tYho~Mxr3!yTU8Lds7_gUz*x4%r$_K+1FHL*Y0y-eM<#xb?_kK{Tej;`G5bowRPFMUf4q!bxjnOeufGk zwMt$D88?mxXy8tIoK4 zmeC-j#ru@Nv^`DXmEPZTCd$y^;4g6rB7CI-SE719AuuoLyMB&@B=&L?NgVto`Re!L1hoYLmJW(GrjJ&6dcHkRlo8>>C7%(W$?nIx}oHP_pQB*3Z~a2ikLMJ!z8v z(l)qTO>UOb-t*FIyuL?oThWV~K(MF;eNjpL340+v5F>=exZQo}W@qIz>vBp=#&S3S zxlw*KR~6PP@8&rk!rMON%*$$jD!2Q#Z^?wqf(?)O^=S_|ig_vf_~?MT1+?kc25F(U zO0*7_TmGUb$bZKG9SD7$w19sGBu3*{M3064q`vQ#Xn)Ht?;pCSs|JY(TF?>RDjXjb zcH|V`@~aVP7wix2Is_;(Pb#6GFEEmgiCOVhbjH9Z@{rQ@8#TTk03zw}I1(6eC~uVa zORO`*ACv-_M)+?6BGX}G|Mp|ALvo`DB?QnA)eD!k zXPxeY@)GBoffN8Wz~E!;RN=N@-Q$0x77wa(1nEA2LuDDAxWViEkaY@unchVDhv4vyXjFJd_%zAdpfA9mf2l|NpB4*?Si(Sq4yagJMEkAZbt-$> zZLBMP3+c4LD4D|gI?KExwWNM!pdZ`T*FG6amvUcE@ZZ%{pk9BdoqR8t!{g7QU_i$Z zAROmNG@y$iDq+L}FR7Yr>#6_xa}g8Fb~&w-~(7ePTwc6^xw$dB#4=UOFkb zYpb^}k61|g>O#+~Xa4R{M!in&TwH6-Wi=c^z8KLcCgaJ3*Ow*JMU3f6T!Y&+)3n}N zum_c-l<&_doE;tqb%_7$YB!`@tg#ClVT3?|Lr7ISqYD7$gdi>6Tk^#MyWU)Fy>)wz{yUz7*sP@-+Sx1Zh9dG9sZR>{rI`&*JyNd5v9?~4n z59I_y0nWjViD2O)#`?h&fA^PztOopgt4BKz-DawAi-9_Oz=p&m^-}WOg{Gv-mizR6 zz@2pzJ<_~f0B~2l5B2aAeMx&NA4}=?d>9O}rykD>4-d5~R9{oAl z=l6p2Ls8HB>U=g%ja-vA77G%DEaJ(0ox(MiNbkfhQE1$dHgimLCqWN9y^hzH0Bs=d3lZKuaPR z*>E|;qd1xXWjpStpeu4t_uy{b*B<=C50l~0I;*DyytFXxv&U$^M=S17LK$R+2j2yX zwBw| z7cutw&99CLa`Oy-04p`!>#ghfO!v(U3DOUM(%=E-?)cRCm-W@%oswHTlZzd9_Uk|L z%6aRc<^?_>sCJP#g-7sPLi{ClhWIg@$&!Z*OT6M$7L2#a{rGqK{<;L_^uOi)4f&f9 zn^tWkY6s7c&qwwm^PVUQp@M9?7$k-x`pDLOMXAgM(U7F7Oe5dhgsx_H7|0uvRHi28*RA=@kAh!@Ym-qgRSoAI6xjPJ&K<_^-D4MsD9U3`baA3q{5xNoi%#zp%gU$S@bnaEhZ;0|j}(i)$@Jeg-Ohc`3{{zf3zbqlDfAR1d}{@{_|bM(%hmk zDQN>aM9h9Syndm`gueXa2uUg^EK+%Kq}0n)A>pV_K0Z+ZR@#6h%9t}fLd0EK1+~|o zB4VO9hv_@Mfd2~V?Bnuk;{zH@R+mpCD%gg5HsuxJT&nnM!RizK02$PBl_z3bk&B8) zq2r+U`Zt}Gy=F6>6(v}ahhcdo$eS3{Q}v$Y?Jeq{EU<>lq!!CS_mJ0j92Qjoa6d4EeF5+t7HOSK$FIzZ$6u%6><51$1)Nt$G)ukyL_Ph%V; zyhnHnx9xY#FKcG?k8~4)Zbu`#F^zWkJRar1EJu8xMI94AeCL~m;)FPhXuJ+H@`A|G z?&jwkO|UayMK}cGOj28+akUQfJbEA4c5E9!8;0x=PumxHnVJp*q!pA>_;n(Awi3bJn&?!uMV?Ol``F#7H}xni zbHEJh!HnnXq|K9rY4H2r>XP@nDrJ2O>e0NqxiW}*Tg-ZvKIhIi@cL}1as_nr%<`lxef{^;h1t`pNagwjZa~J@&46Z-@rAc zd*M@`xs&Nb%kt|TpVp9CcG0T9`}{blzPc~*`Si3DECB8!%nE0L6P=&3(gSw1(a})3 zt^)=XPnM$?R{pjZW5#6pa(#2ox8ww@2KW`v&-qQeO#!Wd`d6&Y;Qnqu_5IR6qcuRj zSGL(jK_x2x4p6Pr^4rLB<_>q`vj2PmU=zMcE)nJ4M__Z_OVDE>@AZ`jT`WJu=X4G8 z@KD&y()dLn%=##{qZs|%(|c?3vh5DHdf&fv(}Yi_bQ-J2Y;-TQhr|mpRoUN()OaJW zv%QwBEMYSy^RF}b$g5Rm?K)NRMeqYyOG7s(bO%G>O2TX%6#df+qc@}$V4DAc=$n3* z;4PVEqNmMlAG4Wx0pf7EU1c~f+6bH7;4Q(&Df{Hu#C1Ga4ZdX6tF4ME)a-+q=q zuiq`?=lK;>x9=gH2fW6U8}R6yH4hW2+ETHtPrxGgGE6MUPu1pC>!4>xhx-+60$HSo zryNx%DCw_M!Y+s&r^^YYm|^;MoGvcn5f+bLtk3(Yxn9A0P5eL=K|nd{AkE!ALVp0i zQKNbpDHQ7w@qH>)VcaYi+$u1SGdF{j+1UK)!>c$0XKjMSu33`$$1(sY{svoF$ob(E zVQkS5==BEuv~OKN8iv+yW{cBBl4>{Pz-PUES`e+6P&~HZEgYK`oUdyt!=gJx%86h5 z7j|Wy+qY(G1BbRr%M?z~g_BdLk{`H(UB`}pUHJ%2Z`t7Jy`07mk32DePQ`abs+nDhdUa0ux0AV`!hlQ6*Gfw3{03((Ti zxF34xqmZfR`OdkDv{CNn%<4_+xaPncX}pdC$vJAqCC}k1va!wy@Uk@~j5jM{HeO9~ z#G&NVaDE1QE`7|=GqY-`&!CwBuO1@PlLsCXbb5Sf7)1+c)N@#sAe8Egjm*it-Q{nf z^yjuUgS~i|3kmaO6S~$~`@qJ6S#YSZIa37ydSTU&p8*8OsqEm;0b#qLVXA;WIH!XB zx6U^|_!}D@k;iIbNhTlcB#qAvmCZPVU-Nf1IUu1P*-h;byHb4`U9Y?AH*|hOXk3L# zmmi-Cuu>Jc-ZZs;(#d`tt;Q$<`2sPM!8hUcPt25fTkJq-Or^YYKLJkJpwvYt?{kA6+}bViQ~$PS8z^|Znl8E}P1a>LEMMn%H79?K%- z)MypRzjrV9)0g@?G>Y5ef~RUs=P*Y%jxEV8E2k>Q5`=q~H!o6812Gy*xyCxC(Lv4G zct@<{z{#v2uWf20a9@F5si0fHPH|Lh$b5VUd1bo@To!gC{YBic+;5cud09^0l@>f= z1aKgMX9mXT2JWka}G)9)oxiZ2GAc;#o-yvT=re< z>x-I+F~a5Tky9({JiEv~A~ol`fY#SY!$3>UbTpNe)WYlzX~2+)W|kb|%0SbNG}!WQ z>_z<`f^;HUx(qCkio0Hip&(9$P#Su1DM zN|RcMa6&p9)fp&oOEmzjCEGnD;Z;GH2kK%la3gz)~Q`-l+ zQvfLf%F0SvinoQd=kuKn?{XvPN8giOke4sjfOX$FjCfz@{(}fi@Wh?*u-E2xeyG$%b@dctU6-4nYv-J!x1^s*1S^OX%L~eYT)%{t zn_l6@!JRIboUqHXu~_{@cYYzLL2=m+_J(Cy2cSoRsPxqyZf7MK`siF6FVxl_w`od zu!;LG6Hx@1>hi60u^l8Oo&FB$h3=90Yh*vMBQkJ7-wOoaAR(7M^k$+wRPOo3+?BZ6 z(sj~$>JcF`R{bC=sj^=<(aYtjtE^=Fs^w;wlsGzb-`@#4OV-3FM6W{%s3io1c+AD^ zbS^F94Tuzwci|K3Ft0<$=>(6(RDZ??j_Bw2_-Fs$;uo3rkpyoR6uY1t3pXL97#t`FL=mgN|o8CViC@O0HOe#(sOx0I9GWL4svz)@~iu<#C`HIbNN)C%o%w{UVqPfon1e`$U>@&8U{yeBwuEE+dq!LyNrt%-ii@vDQWFK zPvIr%hYiHW@AvEb%4^jx);&kV#Jc%|^HkNowIB$A4whht`a7iR>&*beRl>df8nhzl z2_AqcY=mxh8;w+6xI%WuQ}`4Rz1LP_?JLL^9|n@sXTe3cZ@jR<%~?G|8!=)u zI{M?yW;*$Fn0_pGgD}|sl)iMaTL6vVUNUqcdQ^|MeQh%3XFH*F`<-K}0FDmAF(H^h z_?g}A)pCy|P6};hdb-8+FIFj~Y7*W1AFuE->Ik|$arHhnea-3#ZiU9xO0rk8?q1hq zg0U<`Q1g`@ASFSdm$1sYzyQy)aL8YPG4BFGw@qaly8B*FBAq(=svRF=MNI{qzsb5v z&n32T{HLAh2LdKD#}$Ilge4h>g}&QIM|^AV_cA$t_Qm3MGry<9l;4Ttm%*1cA3iXz z4_qf28$B=-oDlD*^*|*QcgLfA9=>7m=RNcm?qDmT8B*C2#hTcB^Vv)@+F{6lt{zsOPu zy7dKB94Wi!GkySZI*5h*VlnyfTa0M=|0xX%GSK#yCI+YAX?@r+I9bE+BxJg#wN-Xb zABlLmpMu4MbARPcYw@SOxIS@e%wJ32w%kC8-r;rsj%zS09{k2%9Yp&xuxZ3`)r92$ zNmVDtXwCh(FRB7TyOAf`ldV9k8h&?qOJYK0bb2#o3nD3t8%-`cl1#5{Oz-`9VcPFQ zftaPu!vfwkwJ%vfeE|r6W1|#&JrSKZQ#A`)eS7m(uMV6v9T|6*PfBX;|a4dhFsnqFMStQ$ zEP*4doH@s&>C^mlC~8@_8y*-ztw`S3BK|oKmQ!pbOf6{D)|JKXXK1Fz>U2?#m%k=h zXXAKMqY0F3U9FHFhmu>Wscd+@HP>tbD^hE5Yu zBUm8%j|bp%s14t-U(T~A6JKF(!pLBL0R*85n>eOT@f-c>>2{lMMM{VIsBE0upjxVK z>UZYXiVp-F1+)rcP5({mbAV0sN}q10-qM`pNLyh7f}GpJ)`vN(=Vxtizk&vQ0#}%J zo&5sX4JSC5j3qE#cttYOFAjB=AT{W7$hFzQA91|E9f3$%Pk^KN_ALA~x?z>@1~~iS z?4Gt-wD{m|d&5&Yg{*tzX_6{QYft~;`w%_i#st6#W&NY-W--MYi@=kc7TnMoLHJk6 z0QDsrn&COICO9nl?DS9XmST{o*vg0g5?g$$KL&yMZd+`!2jIk$+PQ#SA$hOX4V&X- z#$z3c8>HqFtF*s^#8?^7Ud_=)M-W)rR&N1IJ|Rb>IrIH~n882X4@gv-HBl?h@ZRklSG8Yp%c2epf#YT`-E`;UuF%_05ZmrU)zINLo8qxb`VVQ z9Pk}uT9QtjG3DFp?yu8!FVhu9NwXth=5?6@vN8%1Q|hK`+0TjkKUIyL!!BA1lWFAu zSy3zVRh21cyW2ozTxFM)Sao0l@`(dg=A#wBOWB3ae!Uy;3nCqsMTO zS=W*3W9FN*L#QV$T#WZ>TiCs(OBtp4kh@OxyBIk$o%}$_8Ho@BZNCG_Bsf4 zef^g9Zt9Czoy-De0Ka45FN;u1XfPPDymwac4<|f}JIKzbFVr>zezR?L72NsiB(+E) zi^cP?@M3eq%;n+mTEM5tELosif?!=Efn+pTE*Bjt%X14D=U~tkN7j8Qf)RMUbQUs4 zj>>x_)_gUt`gq7xo@R5e=#+M7b>R%iIFngacWzhm7w7?EQ*|A5d++tB0y|~+A zLc3?&Sb4u=eJ&FMQ4Cx@Y;xi(nwggy4^9U_=f`)6Z~+4Wn^)0GphzN{C{d@Tx%i*a zGmsrHD=3UepI8GN*In3z9*xe_nCMhKHgTyE(q{jNJpST~!>X;L>a_|$V3Vn5#nxz|P-LSc94q2gK>ZWIVg|%fVSl*quzXy< z0A&;&kOZ*{E2(=U`Xkgc(J(7-kNy*7)HyBb`?Z_Dez-5;!F9Q5TYR{^N}mTs0D&

J#aMSE&bdH|ttbm$YDog$m`KXr8uP$xJa0 z-zVVVOBNhdw26`ejLQA6;7EBcp7&}BHm2Lsye3fW_fNSPlV=U+R;|Y%&!a6`BcLKi z_A9@()ie_+!_F(&V2XxSsV*(9*Uc|TN?dMW3tpwjcX$&E6+Opw(Ldb3)LQrjaeh^j z*D*g+pL`hD$$}lkcLtprGX`blUhx<-_w4K$kd)CrBvCiB0NKvb);M%5F@|lriL?6L z4H8pO_Ak+8dxRpBED?B5U(=yi)fe(A0TJ#eA((rbLl3mE zX6_Dziw|UQ+i-B*30V!YHIiNkl83fjvmKw^1!1xY$3l!d8efGeXulhu{^_?^Z=}TU zl_s_$xvvpxBo)*7t!e3Rs!k(9FaTE z-V>s=K~cq*Ptz1gfcf~tOx+LkpQZf(yCZoA#v5H+ZR3JO$1$t=(+oz+hjVRXvzzrm zeEYK%2=AlB>f}W@VoBj)KUVJ&@2sWemDX4Aa42DG@`IumU@Ik@y7K`Rj7D>>keiD_ z8qZ&ap=owMEf=Ik_2St|WAyj6({dR?!3jqTrsUJ9_*8bYg$OU3m(MlN92&xnRbG-3 zx-vXE&>_iPec>bFldmuyj>xjk--3LTO3c$exuBv?)FnTbI9x;*-z9@mV76hd6<;K= zyWa7RorL}U)cKc7ABeOJbYhkVqLOT8MB?Z6=99H9j0X@toxFj1XkQ*!-!I?Qx{b)N z-Z@9#P7y)8#C*j-jC_Nfj4RDvnK8|LIs&Y99C~>#Dc#L0lW9Vu}^XvjpB=!ZjTgkUkT{T(s0+l z--12)Rpd+MBpT-iC-*w%X9Yw8#Dma1GII>q@gaARO30M(09gvB2@pVqmlhQMj0))q z8>$hy$Yg*s_|Yxdsl7LZq9IX%U`n^&KT#RP;jAo_)Bt_w+91Su6aw3Kxgc<6{#>-v8y^xn(+Y4vX%~7kg zJwtOOr9tms&d8%Jx<@HnX>>-_99wQqGupL_P1T0=6=`xSJa3)OLU#ke`Ta{@iD}D| zx6>p}%?n_pHLHCOPwkP~#Z)E@`B&4it~Wxim;KyD2D2ohK|*MNVxf&*gS4rLD}Q=IV}_L?8be0tUu zeS(C0$w1}`|MtN z*oy?-$JS(iWq8D(x5Mh*Zv|>8c{oFhw`6eS<*oLHZ~=sDwt?+gY1t8s^b-fqX*%kL zrg|M$Q(V^!9EpwKa7d7g-2P1dtnO2avYcF%Z(-$zHbS~IUZ~_?q1~D1iJIBYRL+zt z#Vs12p3|3&3`XIin*H}Rg;!!WGD7#tY{4^1J(@yT+ejE;aTZSrr-uyr~ z^?0u#)V+W2iRqoDlotW?vf*BNLoWb;mju^kxiMePjUOKFb zxoAV!@S2jR`7uXf?mc8Tidt@H9r7Hdl9i_W#WMv80fA24fSTL5WLrv7si8eBL0E%H z3hpL2Y)IjLt9Bz8DL=pbmw0=owJ9!=)p18luzoveNkV1;+2w}~>}$I(&>nqP<9U*; zg{(}OH?jrs%t&GIW8{R4Gn=adaq?(C) zCNM_ChLbmU8I*l}{DgzOUqn2w>J_s;&;8_aTd&p$$`I)F*gN84(SL9}h!Z}#Qu>Za z>Zh0T(NyUjy3d3euyo1Za-f?!^XeNami>|Uq`HpP%l_r+a2LZxq`DZk6NK5{ z`;7RJzwYcv9&dYzkGcH&i4^ou^o;)gY+MIiP)0)7#YH=RVJg;Ggkd^E+HK)(7m!Jw zW9~RQO_$9-RpCPAEZ2r}W4O|e^Q)M5hK?n8SXD`3j61~w4inH8rS=zxJ-QTZ$negH zDOUz}K=9MIzzo?trJA*{03KgSw1(sgdYQBE{=197iNIGlK|TpH%4VKdFw;Twf$#0j z8*5ettYjN^XUTJ{Tov*`)R6S7-*0ssicN317L(wOiP-+Yo%YJrJ1;aINM5fP^fLNU8E-72rcY@OibE0_5BiiJXLuoR zqXy1VYh5`dwgyB^ATCF+2~C=w$4w~oZ?Q23C;lNz*u8^>_b~8$Ouk3%`kWWi13t5g zjh)YI>>a9QCDs@HHpy57$CKiR=KTxLa{JJG5vUwIe^n<>0RVVa$G22VNe&bSWRf zjdMb1sow^J7A{DFu`Oob&R-FLFq8`<()D5MwTE+k@MxGoO->Expa2o(6H;r4llla+ z!VDUnLr$j)YEX2(RpAqS_`udZ90svndOghfeR6rU7{iIT5~t>ea>?p<4+^8zZ5~D- z?2G{TRoJ-eWG~P^;ps(Q+*Zc=jD$`{@Ce%n*;Gm8XZqXk_HO1u(gMkW@AviMobfgt zJIl0E(O^?jj+{yF~^))qyl1d0mkwL6XJ(FpM-uGX%0fHt9%pI%7|8Ep^&!%_NcdW z_PAqXJ`qlPN&WpUTO(soDL{+4a^6fs=#Qj+Yi{GA+B~m?#8*aceUdq5Bs3QgmF&w= z>BNIW1%G_}i82wCKA3Vugqfu~%LO)+ubOVG=tou0Ulhz5iXPBAuEB`r`+4p+#CxAA zD9hJ`a+DF5scU?+EYe0VmlD|jry`Vix9aST_OP55?(*rMh!G4SDBJKjym;MyXMSIa ztka}Qr5oyn7C~5>Rnz2AXN%<0{D4$-sc^gTEx{Sc`u2&UAG4A2Kpj4!`Klhj1}oi0 zB17QFe$VhjkAD}K$vDmz*BGu*B$OfD&_r`U3_05_5Gp4`jg%Q!ZuraH$UC4vdo?$X zXe8bWA7_>YI@TMwqWZ~NuGnGkM2cRvIMz+tFu(vlNdsp*2E_${OCV9eGHv07T+JE* z-BjKt$J%KOCx-Py=8p{`ZSYKt>OHR#G3{Vdq!Ri~2kp%CASHR5;06;A;c%Z1c@M=f zd>7f(0nU0;9WGP~*iDXF2J5sTG5M$6@7PYkoefCn@BJ_yZ~r8F)%WA`j44h~E(Vv6 z4in;r^6uUs$; zh*y60gm}&2z$hI54LK!f_3AV-<~1!e6jXA`A}IdwabY${O@vCz8Sn%k=Q1UP9n4$# z29##N5t_-4-m>;Qx^#UV%bGjGASdRQN1Gf|gh+hgS6Y__qK%!*AmAJI!{kTP$uf5YiDai`8_|T%t zia=AgRgr+&W>`Fg+8z9Mx6Q$na)m+^V|1ihj~tj6xRvy2Z$YEo+io1W^>f(ty=PSm zd4WSwj@kSX?w>@9cHcRIr)fBTFwvxHLE}J{1n?zDpMknpS1zvrj?lj{JA z?DzimY9nPtm-w_7F!|}Z?ZD)Y0YY+_phOKAC{86{$3NTb(g!rh4I2FMNOM%EfhPQb9Czv93WN}b6}pV2pXBi`I9tTDkM@kaMrmcrxqdsNlonyWM3 z04-A*MBCy2O3A*V$WB$X_i%s9m>J^GrSNatsqS1BBdL|YqOp0l_L;^OEW9tOYl%#p zbc*goXDQuPK(E=WanjiGB80AMLV3^s_TA6GBjTU?nxS8f6*rcJJ0{}h`!Abof&wr= zXORT~3YH?UK7;$`zA(YEG28EMKd*bCQ^(d(mNm()#mwTt|!3nIW+BpZS`+Y1^rOI&eI zL~aAWt{D=~XhcG9K4Jv*^sxyNibpALa0wkg{kHInW5!(BOSVDWVgoSP9Z+BypHhN( zo2zE|cEhd+Yzv8B9=+eWiviI@zr`qYwnJdmA5^_Op+2-r19Ld=eHl=dG(nQh+`ct*M6}J#3jc=aox_6*{Zu7yRJ$3f1uR)m5 z>WD;;PBzc{ov6%ftPFs|sYTu9z-dB$@j)H_elVS}jd`befZaX&?W38Ll+ zM_yiEW}EH?27V;?w_v7h51s`bvk9gxq*dn*Ca3i?I(OXlP~S3v2JsYaAuFjWpzhU; z6MJDUcdyE?5efH=$s!&tJzmHXN$4=nil;MmyymaPi*vI>PASA6F!pBICmVj>?f^QT z)ROC74NMRGOMRDz#ll6(=P`8*_r(-n6@;aABoZkLdtw}WZ*)1r3}i|T>^vhlKk&$k zZuQ7nerC^fF5x1I&$jV_>Ds_Z#LEz7O^xg^{)A)Lr!wDmwE>3;rUqiWXWA-1A98-B zz@8zujZW+fEw24tcKb!HtQMO~Av&K463Cc0j zJ8wOz#S$MVuqu4gZLCkh;dF$17Ihain|KU35mZsWnYY_lB4Q z%2OTqylspEUi~pzqBgw2Ej}Wa%Y7D?4W|qKJ0Xb{XI-OjtmEVJ8?Pq#CsB-iXaqw< z!N*WAqa6J6+1=MH;J~|Rc#Vox|FwFqnTf7LW6Ur!Bddy?gc)A(G`FK}8yL$w4d_ZY z1XxRhVsHW0ZL)>=8{JN2gP@i;jwZN|sQR<7wW+m@r-W1^dfG;TYXcwj8Th|?i220t ze%$U5nT+q4arXb^$Mf5rUy`15$lrbMyQ+xlZmd16N&5u%kHf!}j{6WYdW2<|rU4W& zc#xX`kALvi^6ED}-pJz99+|3NKpCnf2V+$n|K8R{_UOR@tr@Qe)!XKKcskD?P z0G?`jmGDS#K#S{-ulBQ%y|tg8_laMp1;7t{!M`H6hc`kWD7mj_UQ`gChgbANGpqbv z%diE=C&BlD-7wVVt;=miVbWqDCiyP&m-A&RZtpdx zZ{}QL;>?wuM#uux^hg;!=C=Ywo03mwY^#l{C&y{F?;J=>;mcySz4>OH)NnzF%qE6n z;bM1t;;1nB;b7e;4#e&Z4Y}j-v_W));;ao1Ga0KfuD2`U&^e~Q+(fA3J{=s z^_su7lk6?;R;ytx4JMC!o%2g^s&4rqSn#;_ve8{er?S1mR~boSE+!ZVcGG)E{$+oU zVzgFL$mci*bP%r&*|W510IbEB z0|D?p$oF+PaVBPJwNIcG$&gKp`1U4xq?#%~Pw8!RF4zg%RtdghU}2q##)0FGHcd533YsUlF_6x zkj`A+Z6q(po+Lo6Tqpgr0M_yB9em&6l72qHExyNm_if4`tK;y@&2r(A(X}*GU$+%R zd4RjVpYjKOhyE6(DiqCERAT4`yb4rXWpgJ?u!wO&sCF@jL(~qsL=ppSvek)?{E6g~ z^XeBv;*<;ON~Rjdw#|tbRKPCgmXJY0w`p2)qO$w7SGfZ|ou4PZ+W_HA1d#BZm)x9L z`-ThCoQJGCLK60%2l z8z;-W{QSD@6(c;NsV?_=yZHC>c=hXzD-h2)2oC&KJxs>;t>4i8h`|P5bI7shCu65J z>LQ+(Pl4?B+4S`}r7(lX?{C8;Z3)>oM#k_;bXGZAaBMFnI=(ONDM+19ap{>0q{k)Z z$V(##D%xBN_lmw01J&PKIzF`p0wDQ)$x(SL;Y(*#$A)$P>Nu??e9XfeFi>O?ZG$sl z?LzbJ%K=TXV34pDm|P)hGLS@n7K?Ae7;Sx4lYLv%arkdqOgz(H+6q1Q$s|cIE1HvnL_5!*bDfK`x@&k%t@3s;!pgM2I%!vN<#^Xg`Pe z{;=!i5LsFPmw~5D5(seteixA{Lw~s7wl1ln6XM5VT8fHukPFX0bo4nPlT^9yqdLJz z`jVIMQVhMt1Q|FVV%V5xbSFDVJ+fVXgyC6|+z=ShYJrVG7-U(t{g>&)7e&4=gd>|v zG8S$g|KEWvg&=_(z0QnjfG|0GY|09U1Ib^2w1GbPd-G}?HeEvEOAcqLt?C%^f<%;k zlb_K^3F-h}#*PO~f^iV^WX`+2G9$>5+Ki0KXgipFy~ECJ3#m=ZG4aG*>?@3 z$`(`fg~O>C12igSctF3f20Vd0_Cf)lWrrhYcLn-pLed7Dq`yzuDjX0;Idr&BB5;1l z8j_G~1mw|?Kh_fDutX%aTH_(EBCiXfc^Lx4Y>XJlO6WB6CLBa~I*o7%n7~rOdMEYD>*do4`;&VNIn1ru~#K6 zc!oX~Ow0nHF+_;5AQoj#)@UD@2SRDhXGK${t*939tLx1ZQ$ zbTw#_K4K}T_j>$;FB!%e@{DmV>m6u2K*hrC-3O9Q>%HEqZS1!^b;%e!2lMv>%+tR> zH`O=;a$q45=MIVjYx&q3f|IBD<$RJSHA6}ddMuc%-a2*qi^qQEV*p#HaoE(C|`k)lZ0%Pg3Nctr`Dk$9Q{!rN!% zMi=)E^vV5b#7mphF*x+}%i+3??(T1(0ddl`f>%0&*EmIMH}LioloeFx1rQkXu<=KD2p}SHo!&*N z8xYn_=y8jt#?A3p0dbbp&sXUlE>EK;Q&22y)9$03Kq+Us-F&EW=F0Md2!CDZfsY|* zpt%vZsXrf&DqnhZ^wTLJNf#*qLaRSNUlB8H7urXM$HkhvIL9`*_Nd8#oUW-tA?Q0N z2wVf>5(k7LsfZ#86TM%SY{yV(T7S-U%>N2Q{>y+KkQ)=qd2hnpaz6CO_hVh?H-{Ki z)zrlo@>DqF5u%>u-vrnD2c=ffi{sm(sK3p_&PasC?i(fd`|O1wKK2JK;n{mQVJhe7 z0=dU&ZzSj*cm;bO3Hvt+%N@VSiyz$DtVX2t$-6hQmvJ_Ew!^6nSMM*{#zS>3O@S zZ|yJ~)D;Dm+}`9GD=9jR^8MVhDxkS!XcFmgudY&*gIOwK)%46X`erYmOKpqMkS)+7 zHjIP?JXz4fYi(nrdbsvZ&V*xEX5oy6`TnBC5HgVY(m{NI|2q)x=0>Q zs%m2{wmsdJMJ50Y)*|y#0ca5T6v|jtDu6D@pA?9D1vQ3Do031Sq&`vMfSbrhbez&gx zpTyc@Kuc*WsNkYI3DlG1G(rjMUN?>ZnzTIFX!HK zJt!aw?vN~!epje`Mn0{rwLq4RhTQelKGNT%6ZA#bykJ2)%?UV3J}(f{>C3yPZZTl) zrB>T|2(75R5Zw9duGTJO+b-9Y8sF{&Dpf|#tVLdsSYXQzLPC%fp%wM$7#Emg92+?Z z-X`I_bNaN?j6=Y3rWWkmTEWAmFQ&@Od#5bxquGAOo5V<4b`&1 zpS$OfIN+gEvONYJinMM3cVS=w6atQ+MrnC-9%$L%wNH$V3t*{3BoyFPc#=9zedwsS zui(&sDH&AVd(5vNt?mtvU<*OI*F|u3G3kOi!LQ9m~_QEKq0y6ohS|WB(QRbP%&-E;LrMe5j9o^MYoUe4 z7NHjTJ43cB2~M8Wpp-@w0lv{6q(#`@M3@K%AEPzsKmYzTq_>t{=Rd}{OocX_F)7Sqo9nu==W+B?L&vt5NZ*GAL9UDou>fH zfFFeZ|LYzW9@9m61@}OBUktOrd$lDbhJ!pWd0!S@~sqIynn=+Ry&KHe}doT zwTFC}U*@2k0jdP&-?2z{@b5cP1=?_#YeMzDUjh3QJ;XP`h_?E61E44Zq^>P$I5d5B3M<_#W1jnT)8_a1*=0B=H+1wXHH6SwmdP`vNyw5@Sc^1%Wx&Bz`6lIsB zHc{{slM2$(cC6zA0I0`-z*C{w&59}Fd$vMyUK{w0A~2MG!I6F3_v%jR17ur?vD17l ztC7PlEV7opJ&F%jbHWRO!@K(FZ|~42iYjOUP$Yw(@Fr5psP0F{%A5evaWgs5!-ft| z=}caqoua>b8u0+E6)#nTYSN~O$&xCpZ|)(v4-=-j*9R9c*<`{FG2@x2YH2(?RR05T zK_tXqF;G1nPl}wvcF>Em{>S}1cV>zl_5s5p|YKU`3n8}Bg8;1}IgB%wkTO<^kjeH^~zhgK+J)y_fP38mH-bcH~Cxo`Yxu!L7(0Y+%I4QKkT@_{fuu_KBH0- zKokOT1%sIu)(8hFc*kRa`JFO0P>9_k5qf5Oyn0B5_*{kPNL`-+J35CLP=F_S;?ayU z&fw>%xYrv6&tM4N(}ZqN5C1036Z)kr81m$3l3CU3S^=Q7-Mn-((^EkxHeDN{Jd=Pgu%<9Jw(4upvW z@X|VP2$+=9ysf)A36!W zlkS6XG0jT!eutIO)w_DvS zAyytfeo|-HkHt&?H#C)j@-Pqg^&U0n8KjBQe|4nmCGbBXun`q8tjOsxJs_p=XJl*8 zbB5oBCo0P2-yik5p&my)_;f<2*$-|nCq? z;_0gs4%SfQHr@=?{6pOGGY@v%wsyAwzXBZD zL;c87J9q?!Q<)D)WG#^Po_N~mj-K!I_7*Vnm!ZCndLvAEH01CgA#%AAI)}K(h$KO) zJ6?dutmTK9?8AM=AQSj5x2?*Y{@ibIMaS)a8Q^qI=b{=si7e=XTE4s?-UE~ShYQ8t zj?x@VwN;S#)m*bVhCL{aZH)AVw-&f5rQ(jLxKw#R$nFW|sNWzjx}i>X;24HgzIEiq z#Yf@&925f8d?Y4}bp0bC4HaSagZ#>HbdZOMfl=#5ULpguYajS^HH<($Up;nzgUCSR zi}Q#{4DZSE?6B`>+502(7@5s!^?3H-X=$lAA(kt&zn5Ey0dTp&ItcbvGB3BU4h7iw z377;)_PfJjNg0*mF6FiCN8enN1lVIALqTnTkYx3uB$v& z@2V01NIJ7EMY$yizY-{*UNXxh0y4-QqRhxV`ScxBuhVD!-8ER2!rnVGBfj7;;xB&9 zXVfm|XAVhDi-f>3tTZkj5y*A;iHFIm=*+ZSYpMAipZhW_E!ZYVAlFBUJ&OwgcO71g zgN4+swij>o2F_KIi>z+UI8XM+^K1{vAQyb?tAr_B{T&haidb%ARCxa2F;G}u3LoEt z)SGh_RaTJ00ShO-g~&kf$v16-J@^}3Rd>1`@i6!!`{G!>l$&i$u!w-q^An${x{}hR zP839?`D80I0CR=LcVS@R9@^LnOYZGozgtgHooqhOrvA0c24a*AW3RLRxg~#;D!Zm) z@MLdW_H>q{d|@Hy(025xQQ?e{Lw&=d%t&q$#zi~ZxwM@;P!h8hd+e$LC-+qP&~ zT~9b!y+7#Y6IE6FnJPg0su72~M{gmf@$^inWF=$V7f1tsSEA-|g>~)Zu8 zQ&85X*7F#N8uCTL4hNL|tuycUvepr%iD*|IiOBEnKw|V2Uiv_wm^Q{V&uGB_t8bG3 z1~t8S{bDgR-Pa3cZ(SmBQ;haj)fZi8IPzAcd<1(5&0cUCac*XlS*4buFZ^n?NX%v- zI;`})md@{!ql>BT6iPvR(L4kQ{%Zf4K4?DN)Leg-*WtV-`-OxgqHWD5i@{*~=-eLM z7g%i=*()0dZb|6hb6aq6OqFLx=_kO2w-q|+t6O!4d+7w-R)2Hi`CgIV!^LHodV=3e z3&?a|=vCg#XxLz7BZ1KwKNci-%?0I219d#y`wNHJ2(ZnxpU?w&Dc>h5p|Il88<*2m zX_uOY)Vc=^+LIN}o+@cb)?(=2OLAU8<;Rn&#p)d|$X#oNtBsME5*|NP2b9-2v=KN~ zi=2=H;=G&+xa>7pk>!nux#Mz27_`bdlB`MUb?coN$5B?~j-cRQU{uQBA47DJ^Zu-a zGPSX-B{hh&92(pv@BadqlBxF*l1Rh6Lk!D7&J5Od<+z8Z=TiLZI}5sJchkhZ(hkgI z7(MVx8)fpB1k|J*NDPlZyV+TI6Qgb%XU03KE+ z9%M(zX3p5H1UMo{i4*T5Z=M_W1iAP9Bgrgx%Vqs)D5#|^mzO(D3@fum`o*|l-#2iQ zcz*KmbT=l0&7#fi9mVE6uM!OYhWR7J2b_KlAM3H@2-!z-6MKPXtY~fWm8z2p2aO8< z>UNsS$lea<;Zsn5Vmle4Zf77ukf7TOM2iF_z#o#j}NP5-|Wg8cxOW5&CK8 zi7>HOkb)My`cxGkrL(9#g_%7zI7vi9i|mFB#`(=h;0*gk!*{;}?Y!BPas)2NhdVB~ zU2<)F`Ll9Z1wo^p4Nyr}w1aRodTo;tbszLTzomefP}`ENKCz1_5=kTjy0Ekp?X~o1 zg^Et0J5jI4_jGu{c}*ZDzy=gC<E&QT$)4ufK^YVl&UvQa2 zBYu_rO;-U{WIL=q`1G^Dv55PcKZcypHe7){Dlrlv_jP(YRuB*9v1P%4VX^GtdVd^S zikjrOA{L?qSWzeChW44yi=t)w_d2`_C=rl5?V+((77AxuIx&3Eg4@sD_`~x3`h2>) z^3o86#>UK1rhA86or%XT3T-RX|A833JOcICy_ci0Z zh_Id;O+Gs23;^+ySdgc#^HVNzBrCyRy*D9qs+unlpVbxC3FRf>?zpOit;|8BecrJ|EQO8K1N+TQ~H> zC(KTEMU z(GOD4K%@5^<(7P=Nfq@|dupMrPYy@oEQi+^5k3XbgBLL;Lw1e%ioAg?g9fF@c6;k? z;TzJVl7$Ys*wDK1_(ilXTY|I{A+JHOT9mM&dz;neuzxj&I~vY#{PhR8#<1nL%T+*W z(7%H6ig27kH24!K8S_j$FBSn_SR@!xK|>hyJz(b1Q?!%Ij$J2-%U zC$%i4FJG}ag`U<(qPIe!&xSejb==GOOK>!QYm3ZUXCz>-7bbjgVG{4>Oll51y@Hgg z0C4d?RpLk5Jb4!C_;B+#5F0_{_c07j$xgm*S$?_SPH`K-uTM)CsBvnAd-79yyNhT_ zUL}Nat%`nmRDn`;x}u^bpV8K$kKncl(-Fc7h*u5aSRTnBNRe+e{i_UO3RwN784+y% zP@LPkBdO?WY58Mx{p)>i78^P|^^k`Ix(GPYoia6ly$;VI?>1UQ)G5(*ibG{hT)N@t zkdQ0J;{8dTnL>lidl-*-Q)a!lj^IpVan&jvuMxZ>@I;u^d3hJZ61M9F|*Waf9I-S{f-rvE(3qH7i_29w<{a#R;@(qaFbHKXR5si#P z9UM=|lSMmD0l-CeN%cd`C(ulw@!uD>h0Q02f~$&uR%zMhgZ25|`QCs~eNv{ECL|BZ z2!DNhZJm!{cw*?bpTE4&iTpf?!}|;-XS;MEK$+|Cfo8Ao3+2I>D$*-l1&Xf@NyoT^ zr5*<*3W3J><*BoammGr^L7d*m40w?nMx1z+8-ebVa7de!tHhVah11_vEla;|s4bm!2c8qXt_+Uob`27dWx|kB9m$ z`}G*!^$+&*q-qPgVlaVOe8bCDpypd98j^^n3$B59vP0~GepXqNb_5WIh{%{6!hP?s zrKEii2WpnTItXSwARbHzNjp~(Ilud4fEQDW*17j&F&3^Ay^EFLVQ^ z6Y~V!XG=oZk@=%-=aMVdm|dAr&O3F_%r+U<6I(}Z4<{NexH03|fL|XX+@mpsDIgUq9dP}z&*voqpMaWs|d}j;t z#4Co6b4Rl&$!zmis8AHOS}KJO0Pxc*+L!AG#?nV{YLL~Z*s_LMZ944t2SY-q%T51{ z)3mS=q4q2d4AZNl!i`j6Pks|~yHm3Y;Z_QZt5bd5z^pco5{~u({Vqv%zb_0{@V5$p ziQ;s-E;uF8Nj8|Om@Ut!2mB1U+1dZLxbOw-2Sc!hsSwUzkTEHmucI}6We7S2VL3ol z;ImZs7Tdyi;?bm^(yZHPf4i}Z#9LcU)gKpfdT0gA0GxJ_oTs9k)dZ4*HHyV2O;}Wl zq{e4q$hzB-{Udl%&$-w@Amj}3T_eJ*MHXigbbjBM2>;ST*uWiHO^sQBwjiPGj=q~ z%VTqbYcuslE^{)3BIxkh>1~%caP7)TUmUF^tFk2=9$@xx>71+Z%X4;jvrp}CD!qJ? zp}3YY59@@iaCtgHVR0MEr$;3xOH|O!|8b!03^$>pZ<}2pFNc1A`@3+yJ;~7Q!XJW1 zPB6ym@5KyBmQ!217#yRsXcUL<(o;JLUFzCt{AazTnyIm+QR9uqu%Z*d2xZme#k^TR z;onW#z8@N^D0*~rE~5<7#e`m;Rt%Qo@LCIPA}HU=fHl+dH`Go)@BTTh$QEEM`7;@k zf;PSjM=H4U@Ixaz{(N{&EudwsmAW{>!&Gm=4x{qh&3+dhdGV$9$v`s{kPqY33Af5` zk*qUBS!M;3R)cytY2ETgxLspk5n+2mD8{IneuOL_2whq$B*x|tGW)jkZ4M_&`{jge zw62OiI0ibn+sb>3hS6%{WbeOCF1FB+e7bn7;p(~Mfgwm@I0)kl(lIhLa?N>o_8Z2) zxo!7jhqbABqn$_J~`GAESX~Eidg+3-is{&_w?t>Mk z3MGDxr-l4uwT=r)<5jwh~^;toER)vpS@u^ zjRNd>Q263B;g(<)cQ?!6+To~!e}d2NB+>r)EP$TIND+PzI6b!c{l!~|%X))_t}@Pu zX~35n)mj`}6}m*nIEk#mj71xK$j|(ckY5BTgsO+%#46(>eM7P$C?MlfKdKTTiK$|B zXkVBPJ_KLk9R~Mmrwfnj%n9`b$qnx&ogJtZQDXppD^yjvYtR z#`?{MA>T27sZ)=@kgCmvKOFn!`Yfr=$bk*9uNdMZ2hL;`^vC`>3BW{A@3STI0~owY zKKwGhy$S6}ML%dyLCHpL`JLx>i^35-6&oXRUagcI&iY49_Dxsqcfyg4U;948y8)c& z>5OC{C&)^Uymtt*px`r>L(@$4qT_?>$F1`8^1dqSoNi7W@NBtU;K!bxgrWI z;7#La*KTLbvTq+h#1-n}H5&D}0QbMAxLnfW@}s$Pv*N6E|4jS{+eCFe*T{)A6B7+W z<#?}7_{F|R^2&2W$oVi|x+c4VMwaN$)05!$ImzBJMaR;Um9K{I29+ZYmt@YTnYvL0 zvS2U|@v(XBQ!JI3-Sdb9-r@)oaal8raXF+pI9l}05&&;s5?wqZDhD_OP^pI>`RS!G zp;y&B-|EmxwG6PFb$8t5&_ST3qvThNi8Q%N6a$?b^aS)Ef~GZ+tt~-=vCa+GfH@sJ zAcK8OcaxM8LG0pu^{n=Ox0w50r0Jy9^J&~U;I`X9lQG_HU;jP|8pJQ|Od+Uozh-)5 zAo-I}2%hxYP7gao>L@GJ!&^K5XaKRUkV0}CEb*F&4kfG7=Pv@-?X6U!GPAo9*(tnD zANtD!^a|fkdwwXwGRqwYaZB3^DeReX5kL$LIheEc0rG&GhEyJ=uCu+%0Z_+&$lMOPLZO>Y*%r$7?sPpI2SIwrZ;{o0IJv>%*0BgRI+F7YJ;|76#(on2LHz-M z?z{Mb@%23mz_!dQ4>&%eIny{3?oqQlE6Q8TJh9wzx57QT#SK2(U)mQ3xEwxLEORk( zeF1GS=bY|4e3@=S-Vi-%$D(HhiNT|mc_(Rj=0RaUMFbE`2=&qIz6iJYahDOPSIQug z9&QMxz|AOChXm~cB22Dl=JmyR9ncfORc{ZxON9-W!Rgkce~wfP3{tTWA&3t@bs^E+ zx13dReFr|dmzfyA)|j!4))q!Bfj|GzDRHp-0w*{ip~ir;&R%*$6G6Vl!vyok7q`3; z0BWGED>Qf?l0kASW3GnkAn>vk=%JOgFJ4EcLPyo~u7&l()R9LA>J-k-p%31jzwBZ; z^pQ&aD({+FWc5&h9;(1AXSLfH7MVFnz7mWV9Y&}qk41$7Pi{}IGKP}LCS$2k9t^WnWcyEfJ{5#Gh7Oignp zUoh>eqGThb&!IUanf_ia4fZ^?0q%W&E(oS$r$s>5t@#88(>a&<<5ZY{S-SbUx7Ulu z<|1^Yz}Mr=(kGyb?wph$8j43=n)8R4n#Z#?r^J1f5}Inef6K@#mS$gkJC!J_>g^_{c8|x6@|oM zb$`2kdw|0`hfm{rauAsW#Jn>r+xbBiYpvdt%GA!Pr#EV-r94^~?oQz2bhrrDI(cXp z0Fuhax$4;ydY4or4*2Z1$oz+@IW8Jzd-<|&3JXTv6e6(I_Xj|D8&Nn<%g9`$&;ZMO z_#9T|<0};k*0=K{Znn^Jo#S&@8=m6`>fZMmtl%!?j#^*gJ!xh1F6#p4lIxc|Uw_vt z#4uk8*e5*03G3G8o|z8^QsaKl!DfeN(8}Z|RNwcoQT#A|zUz_G=GtPD`u(#n&I3D` zwZJi@WEo%7@O%s%pi4C|2N&ekFcj!2Yvs<@Q;0F7PH3JP?JIrqe&DFWee=%L+Q&CT z1myz9-_7ouNQf5tiB|vvv>mh%8A+|}FUGhp0IT^B6Rj{otfdfZBljjHjPYyk<)(__ z+FGT(8HA@Va)2e!K@z|okA%F7@vC~9uL;YRfUaN<{(3=$O(IC4xQWf6aqqOm50GTU zDiRzncG5iD3i>l5UL5p-oO8g-{hSly{{OIpaHR8A@z2!4RpQ}|#yklSeV2P#B+TDO z6TJ~EHpbmhc9&j>u7*v&$R0Za2d8_zHzFj^5q_x9QX8#7Jym1zR5Hi;ch#yH2CuFC z02a@x+DmcE9`3QdbOliwQ^%ZyDLLj%^Q;$Z6CvzcesKAGY(mfTQpN-4`n!?jReBC^ zesch4JOtc7((G<|oyOA4oX-eh!Q9yal%|AVVPnMo9eGE=QPv6>`hN|@_6ifowij$Q z5Q1uBxtW8`!1Z}gTqt1f*xMT9U;>@MX8Y0e{&J4=64a>FE6WTN%PqxVJ}=j~loPgo zVRE;1@72nG%;AIDU1A8r_U7n>cgN}O&2uuktF3htQf+^YKpgLZXd`m;4V@GCZNDkm zpUZ|+Xeq>B&J+THe~3;UfwU;i5e67dLu#F|>$L1`77W+1=CMHY^bkCdsP}Xm31v{acU2c2TqO2Biu)3TY?leatcV?a zCyR}+IU5mwNEPpb(9EvIFGiRIbm6gr|M!qDwD;9r30v$=O+!Hif4p1+d!6%$yIarU z$!i&;^{Hm*v+L7w{yS&s2j1QVrIpjdkbL1@3{xIyV8t-2hZuBfZ3ss?SV7+1kyOz- z{TrtY2>9HdCFdNR5qa@1@x#H*diSynP6Y>UGQ&d<`kTTpiUyCM+@fhyJQ9 zJ>-<%x1N>N0eW)o2Lj)jRZWfXpWr4y;`X|z`9GE&b9({%snoW#(0GtRX&DZ3lrt9= z8r#lgTREoT%SR@FK0*xjJtMOXON#P?E=EzHa)PLMe}8Xh&GinSbfuM6L6}{~S@9|$ zQaVae(Adl?KWr}0IiOKUJA<(g@*%bdNFSX! zCOUohm8T|+AElLhVH}q}{Z1kcJn8O+rCusx+P2aHeOyA{mjY>GC$Uveu)|62qw07* z0_^NB@q6TY1TtsUz8g8$$GII2IIX2uOo}_jY}eQ6p?$SsVyRf{w8-GKiqTvLP+8vR zmKtfPO7(pZH*h-ybw6;Q3!K=d{=pDaaCZxqQAVKh4~i8psubKDSPc%B93d{IWGnDm zVl0L9OpTL{%|)Lqhx++a&~|8|b)1YTCuWL4h0B(C-mh|L7BC4kuGnwL?&1{uk88)E zKJgV-l%49yVQtYHpRz&`IMZqV>0m5!&V+P9&3NB@`g6T#BZj?I_ULc#;CMHfBNiYD zQO#uixO`O~3`)Kck6i0=D~QJgVum>MdhA0~VQ3rfQVp?Quse50tr_ub_I2QlgPz=w zN$^VWuiip|%Kf#%8Y|=Q-=XJR`zqS86!dz)p{M@bxm+rp9l7Z(!%Ac{E zDFcko^P75HQ|Ij5pbPPlJXy==55cys;og;eU8EOecXL#@DV(ZwQ4S99X$E1lz{vy0 z;wPWJd+&Bf`d}g1Urj8DPjEiWN`3e(aNPb(7`H`D>saC*EGwgGrCODc=4?4Y-GIF< z(cx%-`}TWg_yv@Y2)92P&w%|)O=03f_S|=VSpfig8pq>RiN))}CLid;q@VSbPj{Ur zL<)@1b99#N4#*>fS|L}f`i5%!^)phK-a!ZH&0WXihapxK2XV=UMY}$b;0yaiKLe<` zZU6BX;0ooa>qsQ`_4Z{56qi7%`wIDdl;;|AluobnF-|EJp@=vnBgql_V68n{jl6u1 zpELi4FrCR44-@(Ho}udAhhp8c)6+Lv^$=(yhyX*%mya~0y2VcYk5c4r&iesA|_r zzoeQ}J%v7>Kf!td^Y|iC#Z9&Rx`9mlH`I`uK76vqK6_n|Km{K$o(GK{0^XSQyC-bbbMcHk6XMuHcyJN0pagf^i8M$nnzq~C@do+*f_CA4nHkbF~fP>ZM zk21BNSr$RL0!$XSHDOz@JHhE_bJ zLSW92;`hR@7{`ij^?MGdbcBhG!hau9q8CgU{hSMq=KB{Z${E^&Homc3^8+;%?WMz* z-VOca`A-9Crl9$vu43MTR01RbSl6)r_Fd_URLN!jz+nXz=yzJ9nlFk%Iv@pvptafR z9Qw36@EgAIQ<{M^E%pK*NCM@B!&x5x@~(D~Akq<0M-q(K_^QfzCHZkU$U3enpi<~} zz+9|=JP-Z_DU3d3EXt$g^m5;v-i>>@QJ>~|+YpLS%SOXPGlhsJ=(fdp;8h+_3ZyxA zr8~#CNO19-kT8~OEbvv+-f+lh;5qvN;Vj#Dv;nh(olaC8t=khiI9gQ@^LDhq$5j+H ze|TS9>+1!%_Ha{ge5bF;mk|8NA`Z2DMwFHHNW4*5WXV}NS6}sS>!d=p4}U}s zBwJN5PfWDC(h%Z#R0ig5SWoJ$8J(o|pN5p|$7k$LBK4Nt=48n>AgSU0OwadigBT3p zcSrNd{Yfl3&1u!%0k;czFxrZ|kd@GnjRfW{=O5rOAP$HI@ zDLOfYqnrniGBI|6=t;gL%%!+ml;RH3BSk3KgE$Y6D^!s^tl?L9;kMOM#3chSoH za9NVcF0AuCl|*DgLVJ{wN9jw>znNYv5FHzBV}DWPf9UHtt|`fGGM{GYTN^IheIsuN zxq4Esy+pGsT3PRFrCMrV#5bpGRPm;ua)MVTDUZ!*E^a#YVBHk+_Ifqb^F~F5A00Ei zxJaHg?uZe0nk|O^6V`|qg6Pi&;8Byot!1XWYS$icaDc$ipT}Hy*9|I;UyT9f1Mz4Y zOaN#jcN0JH+8rtZAQu;Pn=kPZiVBCAYt zgZWK``2|NvPz&SvG%yM2LvgmYQe|cIVt>K*-unAEZX{5{`bE;~IkKP~#tVSce2QIP znyC5#a8l}Y$P5ki~&7T(2s@vK4T|hWb#7bJOD4%h~M1^5ipuJvFUgV3r^;S zcefN?A%qtDF;4hs&K#~%liK%-K%=F09>-1)@2k2X3he9Z)C*LL-;KZtfFXt}LS^$3 z-e`WH#szu@X1xHEdsGAwcZ^imMZtOzZyidB?c$CC7=A<8nVqOAV$-nl(hg<63@{ z&{SWAV_hr0pk+Z&qt4~EeAQ@9A#dJ(ryHiuN&#ubVc-T3w^ulTSJY z-7)A$*8T%mtXCCfws*dNn?Y9mU#=MpU|OKXUCVcL)MD{+w|rA1zR~@@H#n9;Fj719 zxZtfmZA9CP(G`Rwjl30??-J<1qWg1gGY_?m-VU4D*!Uq3hGT+++x42v5YJx*uZ%sb zla;94yX;Al~H&kQ9;b`;nyY`!^n!P$=<& z0J`s&M_#nB=c9VvUie^I94Ce+%vdM2Y}D%ZGIp2Tr|Onp563rN>mwOQejgzS;Yfjx z{rA3E(8d}_ncqtKH`UYOaAKKANJOw7Re;4yK@+hl+`7-W@7g*^Yj4m<#SYAX8$s8f z6Q0u^`N?5`BSQKE;&Hd#UH5pD?4J^~J+Ncd8n)|7zvX9Pd0>+{z+jj)9&Ub+9Zpyl zWaT{MP~N3Te*p6|At9Ql0cHJu50vD|w12HKPxR0SLH=3P-Xihq0GgKNoTYmhorL=Q zyOS8m4;KU>=>e|ypo5|o57ttrkB{gVrO>RhBol~lo`d;1F^8C;G8P@Kz2%fU!oCsX z&j)$P@71p!ZL{K^&!B~J1SHVstv~U^_5-44-(~F%>4fceycpus7(7)g1P(o-^;xvz z>k<7XZWsn=j(l#i_N$x5sBRwy zvdWfmqwm*}s{4(o&7?r*cu^45MYJx49u_3~q#iN27~tJA4hr5O$tIB^4quG7gE~+G zY(IhT9;)6jJ&phEqR5YQ*7LW1 zy=Tf8LXIgx^xw1-M&Jl8zO0{vs2%n0efuDJu{8_$rw(&FasE3sekp7GhTF>z_w%m% z{D>|WrwEHqX}+3}D|nX+&}H5tBf9{7aV{Z`?_`dTn8U+34R3*zv+`%@?X`WhW8MpK zc{vxevEMnz$1$o7ki0o-C9u`;rCn;)L_G&7&moK(r8swXm@V6g@k4&WXceDgJcyvI zm*&0e?-XDS<}S+V0EEHLTgx^`0R&)?d0MEmt9rRv3!-EWXe*sCZ`3@kxuiiyVKb2) z2b_XU>cE6jMTQ7J9a=ZwQ?%qw4odnLIK#XJDf*(ad-C6Vm022&j@E_Qh-9>=^& zfA0tEH&0;z->$-SclnOPJv^s2LO2uv`9Udr+y2vmKK8{Cx-ybrS1(EDVG?Zo8ed+j|4jWv*Ehq_>Qz6M%n28k2EO&%WK1}GH~>AO~G(*AypHQ~woe2E*1 z)^R-KafkS$tK{#C_ak}*f(eYcgOZ*HqEqFNc6dG!XP#mwfTnFF?%1t*RK5J&kH0Jr zo$2>FP(+)!+iWU`8xaRZhc|4f!^OmGYbV$RkeGJamuY!owJo<%ON`oZ00!~kw zKXaP{eptOi>+(R%{dBx9$bhuV)$YS)rCYTg(-oS2rbgJ-52oMbZ&^VEWw}+IYr}V? zgPiduE4a~9ceck-r(kYZWEZVGRVeSKx)3UGr=J=5=%oRZs&)IrvDhN-&gou2H9_AGd(&P}k zNq$GEbH3)`d5YMi@&%xX?{s_F<^@`J9`WUFn<7kopbiOhAHkPYWj@=|A8fyOg$(y~ zxrAHMNlSy$^Xk*{E4(iTgTB$XOPs-y^ADwfsmCmo+XHN zfD7U;e9Bz0ic@yd!4cB>mZ6C|x?jUFfZzQ@Wg?LTXJp0>ZhONqhaJQHpn&H8i06EZc9RuW>=&csNp$Jcb<=ZE>bKtYB&G`*J%;<6DPY z2)6}fxzl@aA;m#3wbtDNOa!_2KIP!3HLP~7^;KaHF}h{yNPNJDON};kj4k-IhyyxZ z65SoG`$HfgM7&UZqw{ly-!y5{AhQ_VEhGPWQ|~{~WPwJ)XZ9pb%rFE_t5CLe^f5dp z%^j}nrKsLY2m&rTTutKCD?zArMe4nX46LoJ&{+J*<6f>}mqSSrKW$UbN4by+cltRR zS^|yUvdyX}vz6RPi9y0-lAa$GB`Sz011>Fmqp>mE_X!x>ImV=-ZvZ6&62MF6RTpBO3fim5%PRaOyCAb<6+gV1VI0Fxs8eSvYMg*qN zsxpC{0W-yM+|N4tUvz1L3{n9NJN-OgTjmAtZSH?2B+FkD?F>XXU}umtP) zcu-8j(!%{FVsHKV86Y{1&%Z?R)2?6@;-20=qvY-W=5}8^+WfbB@7#J$Qu&xVnjM@( z0asCP&vyJ{qJirqU```X+e6(J7h`Jncii&R9D?DryCEhjw!ukso(7AABFIy?lpce` znedoY0@@J`?HSaB_>smICY%p^a~0j-$8swg^I9Ge^2eX<2QI1ME2I7`lxmD9064c* zDXE@(FTo-K3VsFcA{sa5*^`LYj2Z$Q1Lke`3W7sBHXC3x|9lS<%zR`&%z68U%yB(5SoGddOyZ9C==bTB9_4@- z1YH_m_W2OQJG9=Z2T*=N&V%5xkT3y0ql?+ZcXClZ14Sq58;?Ck~a!+cB(-|!+~R)`%R6WxoqU(*W^7;7Z3Wa(Xb91p9(qx-63W^MYMV4Yz;ud_T&WT?!eC0-ejRL zui+j*gm*U1mlDgNaSxAo;q?!3pgQpWuyJ^1A-)EX>7V59^FU{S3s*p_&S*`?Ux+xw z3?}T-X4z|e5?t^p>tesJn*psg&sMCK_cR^(&-j?a3<4b&`{ox$8m8-cC;>0(GMD6P zkMF2_O1We=dGVM)4n03kEQt>2Bl^^@aLGZEi9J3Y$_Y4kem!S2VJiMWxEHZP^oAyHC7vdM5G_e5Td4?RDn!Wq70(x__=6G zhCWVcfkHPEG&Jo(f=vcrjNS^mlZ`Q7O{;<^K-8SiLj-COv*TsKLFr)~k&y_-6ogT_ z-eKu;jnM&%=vqqeCd=yzKY!7$IR<`oGB{u6et~7=jNlICf%JMZV&%^F|qr_I%Kg2 zG`|uQL6SbufDE=4Lfks{gZrODXv2PK-Vgir8#-;-c~UumGAlw%PYK&afGiMwAh_Fl zz607oz!`pjDD{jynp8@0NFASsSw&Dpzi%2fs`s~ERwvlHDmYheQm>qJO#sf%n}2b2 zz_%WbXhtA*EsSr<$;h78TTI_F5@0`&sz9LZd%6RT%>1cas1s1Cf*pentzHq+mde+Z zv?+gXC_c2CoAcs_sorv0?I?2ZAkK43E-0(8FFP#R8P4U=p2%g;kgvcxrNE6up z9H|eXv-nmHF@Yr{Fd_Xe*I)b!8=BSL6&5=?ZV^l7HUvHSebDw{R{)<&e-MDm>6SHZ zRpyk_e!u1NKsvu;Z1Q*m6hZR2%B$U&v<{e~5IWHzDA$H;)?ybu8f;JvD^*Jpq!X^6(^>TS29eub_HicF={Pn$S; z0n!DU9up>^>P?K1>2LUzUJ*0}Bxtp~8mf=aul=aVxau1KfG{0`S9?5GF2{}#yp$1h zw{xV#4ObTSdnf{*c-4`(^b%|sIZ>6l4kp}imX)49BWn&Y)m09&t^9JHSo8y=<|9Ud z9ro4o_xQ%oMnA(glAU*E2wxdZR^N^9diO5FBM(_VYcvt6=)H`Tm4>LySl_nu5$7z$ zDd++qRRMBX6W%PK4&~5oE)%8)6?la2sy?||fusBD z&YsN)-gCOmIGsv}FFu~NO*}~_AIK8L(%N6B-TQGI=ri6iw<377ViBXj`%Qy12*3f? z9&5x4VLU0&vF4dyiPzQ_U+26%_9pO|b^pSCm-AkY>^}op?Z+ipu&<^oKv@evAya{7 z-fsbv6FW%{eypXx*bClhKfG-~;eOn=w~l{jGjq8%c)xYQWi=n(z@t3_5}*rJNxKtL z%=Q6FOQZa+=($vh_Lkk(LGY<>MLe(_7lIrC?jYaMtuX>2uVjRa?Thyc0*QLh-TfGd zTo-3c(*!BE=#aNf{F!U_gGLl`;9op~M9@s)>chT9!foR0!B!Hqhl&~p5{F=gb9y2Z zdtWWo@YT8Q-`9OFocvhMNgN{mrp7|*>LB5-#Pxo-XwTphu|4vo1tZ=%BkA~u^)c^W zTcGT@=GQxyN&4)ujePaT3u-WfmL9-n^j|WZrheQwm~5L&M#8a?Qc?q z`_pX7`e_H^hoOU-gC66P;j)#yBEUVhs!*xe-h+c*pt=zmbKNf*9z8Mqc0cIdD|*Ks zNf)>_f*X*j(r4D!wKn$$7LIsJxhqO1eFcRg8I>o3n>!M?`15#B z0iFnWZa%m|@es|Q71MU;qPNd3>Vr={TNd8x_Q(Ak6)D(%YSn`+V^@?5bEAG!wN#u$ z6?g1wpF-sKD_3M4W%veOgt<;;oSzrQ2X-hW*Pii|&56R>a?Q&iOQ2Y)%QSn7osTx@ zt)Px){a%m`XXIg~hR}n)!TUx`Nzg%MJ0!OWr@U~^2zaRlxzHUiM4o~R7scVo%|M#& z+mV21MMhHDM{EKhhh%w7t8k!f(czeht1?$VCkt?bE5R)>_U0t67_ZK$S#7?msbsMU zEDM4oaN-8?jcnsTn?%(-D-2Ly)1jRp=SgzU7x- zOmkB2-Yfd8Q$|7K4Ir;73BKt0* z$1mA0zyRLdm$AK6^?rWT@dKoFokgbKHX_!xXyWw6Us55=_lZPMd08A24=jX>S4$1@~ zNz%w#RkbcZD)De|j|frnfY{HPEUO)He*W<)x5tCqud}{gpMcAONA-lr{@&~}C;GT5 zeP%2}_E4Xm`l+Y+i}fUS&)=p+1S(fb3bH3#NAvql?NK&y`dJNO|FPPK09q0x&o}vS zKr-$JnVIX;U^?j#mDei_MN;#0DwneGgh0$Y_~TxV8p&-QCSuGN?7H`c$aX&^$? zydjvpsepvF@nTcAs3cfsKl7l-SX))kpd6B=HD!m?cleG>*0=8G`|2b2P697U4a!?lyY0`XrEJJlJ81)%1wOAp$E^7x1 zImehhZ1jk9uIX-sFqN7!qjn^B< zyff0op$E(_gyv%SK0gq+F3&Rz4m6CW-TiYw(yhq{>YFs#$LqA0l)~DBWg^;(SYE+c z#YcX7X>FS~;~^tAUgzPnpg6~^0c__^SB0rpE&Zb_pV2AZU#uQU-pHWgZfUTeHLv@b z6Yponxqn$9_z)|SN%XDq13EP!gP9%tFs>4ygng#+@06$ew#w|48v>3qP#6um3ZM$L z#z#xwQ}?^P6r9VOBCqe`we7qc-{VF2m7aQxZ_V|l5KaJ0e7#Bxj>t{=v z0v8_LC>aQ3+hMkyMFi<0I|S*pm+!%_YQyNt$q~UR_#;Q3SyNGn4YaCmI%7ZMtl)KI zsO#Dg;XX85x8YfqDB%U6gr}K`nS1uwkCd25Zk&uN$*Yh*qRpd|4KEWU&x80wbpcrB zV-|X(k+8YI6`bt?!LUDGdI93m!Fl@hoiH(hQl$tn2TC2I*_F)us%_yq%i#vy+h-L6 z;Bqdm69K8IDCwacl!xi}k>wavfl1$xe^69f*FLA?d^%CT75qAF(-a|L$jgh=Z@Wa- zdd!IEX>`TWCDZ~m%Z&6}$kj33fE8n99pME{4t|{DYFQ~c*5`~(Fgmu4;f5Ww(x@rz zj>law|HjvaJQX6%)%0;JMV9sv<&6`X)IZ<;;hiCgUDFnEKeLSzd+(W7FtG=rq@9)@ z4Af-LJ(wL7@o+1-jXTAspie}EGCdzyIVBSu06f^dyw>**3y{az1+<5kh5Qa=X8GaD zx^311jTT}R<>>lq2}P@_%Hi7#WqBEoRdBt0^m zc|wDOao`#&0ug<)S9oHG4xGj8LO%VNBSlfR?9kgt6ZOuV1@+wn zQz?8AaCEHGahlCKsIVHW`tJB%bQxmgu)ZxtF3vcjgF*`7j*|deQ>kaWk7gyfF*yKC ztfwm1EtT$S9g%bQHX=Q{C9x;FlyTWlG5CWTz(tLP9+c+|sfy&3GxFm40`1oQmsW_g z^Bo+qL|!2AY2w@7GwunZ4N#|pg2r12=lulOLZxo$5Wx>^iq2$-Hjsvs{Kdr|^A~^di;EN6x;C^L0GCE)WVuVH2!Z$6^#zlg(tspZ9<9TWa zTt3{%pYw~5(<#nboE9{WOYEdZYA#JTtMc_sV=W7g3g7*B7sAy$2Ku4@Zor>;o0(DjlZuaPE|@~h)*xT}8cTF2g{Rv4+dOd_ z-S()=$sld1U_o^tJCz6Du0gn8vklpBqiKJn$}gDGYm%3*?BL7F9cl-Mg}JA?$pUxx zzsfyEL0#)&J|FV3d=skc{O$u#oD?`21DFGNB2jBL7vY$Fk$dLfS z0n8xhKfFF*5s2O+Xkxv^x4b*u9%{hlo^s27#Px>Xa{_lwSpn|7hTy!y5++XU3D&d& zDTLjCRB2>i^_m3-AG}ca8{YHJ`vT!-YextLf<#E`NkPD`?;xw!s#RTGS*E7Kz2}}2vE#6(D@YFYb#Rmb04rz;T*=sn zr?Ru$z}NW}saNJC$}mVzBq6it!w|0`08MulhUU78R|eaVno+ny>Cs+S>F z4}2TdBZQy8fm$JnPCA8S%~7>CB}FBe+d6-IdoU3FdYt?o4V80nO`{df^6mHX>ZFak zLf7$d!WbE<8R&|SkC!2-4yiKDR;AcvFqZnvg}N_qFsff5tyI(GBLl5}Q!44nui!PFOVWbeHm-p+7!^{ly>E};XEwM57@4H_C z2sF}WMFQ2y3#B9XnAm4@yb?apI$Wy58yd$0(ClNx6VY-jt#HvQPwrg|5S z3Z^g5LL(n#HTPwM`v%e6J*4NjV&?0c+&f$ht58o61Mx@V5_z7&L64I2BHBae6)mEw z^@GFkGYZm>Bg=O7K&4tvBaLJF9XvSx?74Puvr*&|6`vL`km z_K%qY!Co<_%~1c5qeid<6yeK8n5(BXX4$;kG=sm+1-yY@y`n0`Ig9PVa-lgYjq`}V!IwkbqsF8niQ@mNV-4EAJwZu7n|U~F!*NLNBh8))a*8Q#C-e1BSieRHw4(`$3$AkZoq(EDC_<> zUOs8tzqosGtqOS9u*mQ0j8hXE;X zB9z2S5*$upO0EY~>#quqaUnb&Ow0p*BEP&qioQzn(=B?+MKQ)Xd44}( zXn94W|A+tQ!8dq;x+0T&;F@p6?2!`4hVqOk!p?t$-2x?htF_Xx<9{$$MhD{5e#yIR zBF%G8l!aPCdtTnfD0pAO+6de|R|)r=YJ9*z#*@B=*V9;O=N5dmqK>NQXSzHqB8PkM z*xFucB;8WqsXtSDCpC*ObX|VW&OFlRqj3nq@774fR|Rn&Spd;k!Iuh?iXY;Rec9;d zIWQjyNsj6$%CE@v=;xqUoljfn-R<%MSn=Qo26U;+#UM3ucvIOq5tgt(-7wxlqQFee zXC4t8Tw=dr=3T$0EKHPOUJv`N|Aaz?EQblEUu2Jo#83Pvvp;*6{foew;1t0UaCMi( zdBF;|;xoX?#T_xX{q5rT#%t$$F^v|qP$h?M002}hI*!pPBbSVPiwSWO?PLR5dyf0_ z-v2t3d4kjL^%1?KE499m6o`o7;}P+Fpj(}uFNPY<9*^5Z`gu=y;5yO3;e^|BSM_J- zoPzIz+0Ue^mnLUN`xOO5CGj~L-=^!X(OWJAVolFM$sKFRxmj95tk(l%C_DzglGSSt zzYjJWIJaUwbRAC!DtO9RFTHr^J0}Ge9t&Iw z`iH`SkLBig&W%GBtNFg;jZ}@u2(Q)a25CsWYCwUGM4ZF)K&S*F%^<9&n>;@Rm>jZ{ z6n$cZ>JV$6u)DRx!cDi1D`F2%)-8Mvvry#TO#av|V&Yo`6f}tN92fN?c zx(eI#a>;HTchTh8_DNDn{-UE=y~6|!gGl3>d=5s9AV&k`qbiY^{Y7|rzjd?Ba1#jKAsvhf`cH&ioAIH!w%@pOIgvR84<2~b5d?9FG!5c zVwysi6!(;#+LVf9knnHAHC*t1kW#pzcE>9{2JCE~IDDL3Y|L{RNN{kGQ!(s>5~(^O`}_;)ksgvM@kTzJ0HdCY0vd_ncgkH4zWumbx<5f1 zJFTUnINep^L)2X;%wE4>FA*+&c>+NXG9B9)Vk}&S zPS+C=2AKuZQi>&o^Yl-8T!Zf(Jc3XJciYcz^|4gP(~IVv81MZ4T7IXrz2HK`7jbJB zrR)k3dcdBG!@{KkL|&66kS$#Ge(5`FjF|KTSl4G-VM$g}*ZVY^2-Oo)zEIts4zfAN z>&?2OHl>xl6_5byn7VR-@K7W)r$g{=mDBO!JSRkyn!ylC9^MxAVc(9!16+KF7Oh<+ zVD2xQwIa|+;Gqrzk(v?8={DyoG z`^P{CD$Ao`ZuM6zW;k_0cz(|32k*ySrLFr0#D|RYb7ubWHUyJC#wdcz=4Ni}jyQhu#i2Fzs#Ad9$beWL4pPHPec$9jkO$HC}+0!GQ8rAK0pbEA5Vf|N7}Cy zJnEgJf>Yt3l}$DTasd+7br|N$8T~;4NBBeLRs+fbjy#L+T!W?@h)@JoMHPChpVt+O zvV8~|jJOv_azztE5m=gI7hbijt1f%Y;!=tre4`6Rmv z4hH;@v#-|%H7?%3GkMG=cuZe>jHD@i<@Xp&Hb&Q(ha(8XjeA74CEvUuTWxRv z5hL|d9cV;RN4mr=2gyH5@?N~0sVZwnwUMrF09geeI#9y1g(DWnO-tH~9wBs%SnrL> z@RGd!GR<_Nbb$`(Ilx|>V8pMUU1LBn175|;7j*);7*W3$-6nPX@WQ=5Jh%gLTciT@ ziL@`u^jLLm<4606RSEZhQ|d6+Y4e$6N9A#v@L=-bPTiD2>a#Q z1#E%_qV7Gu&K84<;gz=(n*x+jq+gDZk-THh`83zZO$*%1y}F&}Wf^N%#OOBTyeETr z7o^m@rvPi(VYO6eQY%Nr4d3L-Up{I06>4+eJEStx9|h0VuHf*5gY>tYJBh@++rc?; z%S^jQ`windf4ik0s6_hG+hlLOKJ8{&7X&*3{F88*5KauAUCQ z4YjYeq262?(%bg>E&T{Ub0U86{oMMIH6UL43l`ab zcs5WK@+e!WTV8>-dtlX-j^M)ozbHJeNly_fq?}DC~(glXSy6vmlAW-<- zrGI3+$I9egN|ZZx|>n$$gI)2Xs`i2Y`; zr4a3I9Q+A;4EJq@D84vZ1yVy`aSpRm4EcI85tz+b!niwkz%=D0sR)d_%F!^ts^9+8 zwAC?DfK!FsP?a=&1EU$8b%4Y1HFQ8p+D?I%L{@8EtS#v))qfXoJ_UJEJT`nJ*5*$mZu+cMk%bUks5~bO_%`yx_cRa8->LW?TCuoc0Oi7T zjh37$Biy}(GSNk%{1>5AZ06%K^ePx7g(gGy*+1?l*-Y*U9`l0RT}QVOX-UsmZ&w_xfz^&TWBTP@YPMSi~-@|)6JYJN_c; zddcmPxmXqh{17@VoPwgO|7mO-VlO-33D7I%?$i+rH?!Nz8NE=$H7>n@ybFq1k429i z=@MCZ%5~D75&k)Zdx{XNjN0^Nx#zV(Bi{vnA8A?nKa1-TFk?EnME3I@6R~f_W~Ndu z6L@Lp3t9`wCb0QO47oWs+PCpeX}>(+2K;RPGWiwNdq$C{A_^J7AILL&UX$TmUO5OT zL&iGmqM|?A5=q_E}-^HYoH#uOk)Pf|&@XPEQ%f&hoPl`WyIgJm+APEr_4ej5U!HXs|J`N0?+O4k-dHYP-)xGI$tb+GPb3 zKJ^zSb+fWvPZ(1Ayn!~GZdmvYX-to4-=tVAnGudOgy}~P0Au^e?o-zvWjZ;K)c=)j zD6rpQe-xhw0MQ%zu$#RDM8Sdj?tcf} zn+oF+BYj&q=B(Qq(^+8f+~F*e$iS0lSrE?=Rsw`R!%R=zF#$wdC}eYlsk=Cy3yqeg zOedV<^*u{3K*hw+gqnT}i&H6*Vfgq3Bq%E|i9hNaB5D#)r#_F8@r?$%KRP`~Mn!K_ zYN5}?tPCo&N3(CYh78r%DkruF!#<4b2Od`~qPPl@7x#(}(0!lK58?qPx&14OFEJTM zNy3lKv>ahwKRS%a@BN)!ALLpzjJr;!{m_4w%WHnyH?Gm`o{AseA&uP{m^#*17yBz> zLFC(0ZuJjnU4@1AW)O&_&Q3AVQCFW>F31q$0#P-s?a729 zZ~_4~eQxe;^UaLNBC#N$*P`QRl;w}4A`DTu9KTm@pfCF!I>!LKvN{?ycET+*j3>L? zn)6H6)1?uwc7f%`d7G!zQJiVlQ9OLMglV>JG6PwB>B)lZ4xb&&xoB@W^HYyMb<~&< zR55hcaI8ExR|FyooWUBNcsZ>nHk)7Zf^;j9db3dY6rZv1=Yp9Y5Cz6ed^JQoGiPp} zFRUU$3#|6@#L5uFLaj{m2e9uC#wl00x2wtddN8!e%-N)I1@;@GIrk0}K`@*nu4e9` z)E)aw!M{h+O#MQ_n^(AkvPzdRW|TF$VRPk(Wxw#t+MUJd*D`&7_lFkgl(zAmWb0awmquJdr}v=nG$(jcb$%@C z9O>guN61j9q5_%`q)`HcgFI_H7=ceqfr7+$f6JuzT=E`VsVnw1w(-J-IiM)SrVb%I zenXPU=Z#Wb@nclwLD@xj|@jNo>}Y%i&VPdJBf(&%CryY$Y~=z~yS`N{#P?oLPh z(<|s-+wstvU$NqiJxrnimL(73r#eo?YQ*%oWdc)l`CQkWe?{k{B!YET8j3-g}JKs zJIx7bQn4vWwk7)M12CVfQLX^O&%-S|I&H3Lc1~!3ng3J)VzmWtn&yqS)<5=T9h?IHgFsWq%h=Z6#dtqdaw6$@-S>G8lzdA0(ATfa_l zvwx8v@2nIG6ko$?v9C~vygBXHjhw?d`X2Be3qH|^)D3aKyoxLZTTL{=sJrl}J%`tC zrK19lj}@Gl4Hk6prI*WzI((sof7DJJSLn&p{?UWph|!zrbUexth^4Ui=@h((E8zqS zD$TwIQr+Gg<*(tL(swp@!n{G8?VTe9(7B%Z-N8T)(P@z){ECzXH6cBu)j-`D#v2O6 zmFxXhAO~lM({We-*-Br5yZ6(Lz79np9{D;^RUY3J^DzC*SZ`tb3w>spvy7Hiv*w*S z!>5d&eLj2wcmF`ZG=e{5=bI-JmF~rfk>u5Hqd+Ok7I@0#L+iefl6gu`A3#6${|gE4s19hW|AZeK+aA=2iE zDMmf~s2(4`Z`ImzZw^Gq`-K#Xs_=ba;4CtxGY#f2PMsHz%dPNE^&mn&c*rL_l7XmZZ+z4lrcsD- z+!}3tVzaa`%>aT~^WaF;m(t_r%L{KdH@AAWlcyypD@t50Ag=+(h4rbfv)_*;)0xDa z2dPdC_%CLlR}dR;qxVWu2TevJLU_f z_T;`k>=aF?T>G0+L9YxOcpS**{7XID3*y1ek3n5XbER$LLyF#Al+hTpAAi8*1RyMI z0~~@GlXv+^_0wCsAx8b>f%Y!#UHs{epJ)2!?o*TtvwWZdn7cLv zT1JM60P`XR2A{Kz#J;mPXbwoe&%UvH#>sw}snwHA+qi8EC4J12%Oip33Q+`o0a?1B z+Jg=aFfdS4z(qH>bC%2^jx3y`;s!L;UiIMDf&i*a3K0e9nO(d*zbf=6S>$v{HnmX_ zL^2!{HpKluR;pg;55XX$K+(gzv2Z~SC_ee*1`_D0o?Rq}fK{9*{7QpkE{)*Yp?O4E z&-OjtqlJG{h3!!;*Y6FF6yQYVW_@5@A4NzkMfNroP37*@rExQ z69jXV*2~&Tig+l${`Lb6UY~}f5Kb% zZMOm1y{nP7Na>7=%#R@DM=3l;GXlyFXViwt3S3QXWH;}7d0+4K{-GgQU=)4idaaCp zGjzs7Ju7pfha%!`K?^h>>*gNfo3^XsBj;K50KFe_{R1#x3FYENKS~SP__5Fv6L4;H z6tdwMh8~92xS$|E{KeEn-Q_&Kw863yFpBA2?UhTf+#_3~qP{20nW1I%LrVbW1+zLeR3 zZzeqhDCXgYpeLga$^Asku)|NT9AY@rfg6y+oje*cjV}lv-b1bn6c4zlkrV+4MlIiv zNmt3pGP|v==X}^(ZtS(nB1{h7&(yd4+RH)5fc7K;lwutWXfyfwmXO*&+>?Yk;UAyJ zd#S6aA7~?g#>aiNO2y$yISbQIaa~$yG7M;TOOC&Tjh_z_rESGt4iUsu6 z-wXQzuzvOadq31a;Lcte8?N;Gp319?>>RI@M1%gaCv$l{N$rNGFY9g{A>{5TC`D_# zn-WxCjHBM0!JB)rf`N;qd*~zOfddbMT0|?e{lP zq#N;zSd#qx1M-~pTo)2#yp)!DNA*W3w-DD_ebh?c_N(a?uKKo482%z@8 z-_d%?f0gvUI0dtA7m64a>qoUP$h@*Tc`E2(Xs`Y7^t%Q;{Q004+*IUwvB#f=7DHE^ z?>trcu^+v{w77vh!>nc&g8Ak4`*?dVx?_Ija0B0q>iDq1 zhytX>D@D@OcWCQFVjC0GsL}BYbrzyrqdJ9l%OjVHED7)zw(x=Uc7!anGr|&gx?(LN zu+T2B{LfXG5Db z=NF5d@bOY)>~uqWkcX{>clapq?!hQffCS)?SWC6!*oQ6iLI?@|0}V>bME8JK3?zl^E-Z*iowbwEL}xBrTx8>Bxez|(nOcc^jPcfm|ZW{)!cW2 zV!vvCpHWQKSFi>uXlL9+NY4a#?(hW))bsTEt<2?VBZUKr?K)&uNWUq2 zR1d88gD>*{fj}BAIC62?xBRR$qC?%la>PIfHldj!%TI?BDVwYHW5h)?5LgV)hYXM> ze__Eh+?PG=9tu)NUF2goYrm^-^QXZx$QnHcE(TbeJ?$k1=Kc_~;IFKk?RcUCYCGlw zjXjn^f}`CLDIlhj$0B{qsco&d6T*~SLB|-_0riFa(q*$O5iSx=vN1ji;GhNw0N;24 z*k zdwtD-)x<;MeLTv=$~68kq4UPgdz20ioI@mr8lQ@y0DT0W6>WwjO)^h^0-=duslbj% z#-0z17%)1nd@YKeC4aw1l2s5e*uPB(>U)oI;R(+^`V_B8eugFsb|ZZ-B}e@pPvdP`U@waf+V_AI5^Wm-oPd?XfxJ#+ngqc+lWd&2 zw4^4|Wa;Y| zkDEa<*H=v{df~It1^^ce9?0wcJ=u^)^xQUG(4S9ly=byQN23FEJ>&gjEqKaI37+jq zZ$!2Qb*G~{-ya_^23vi;76D(V&khmc@ML6R~jWe<|28SrGpp&r%`9rNC-%z72q;|o906bopVQxd>_#3Q_z~5{$c#DnDelH zRh?U=v#_sk`wt`FXT%~mnQam z)J99VL8(Ww%X>QP-wC#{qgu9za5X1&y**4qasiZFsWh9S4?t%+0aGPf3Ii|>*%M1! z6sE9k3qYBGrRFKm8Qs2nH;{32V6LJkaGPG_?|{xEAI~^hcpJ|L@?BQzh_{D2vEK`# zP7vz&fPwWiUmOgGW9S-}W+r*seFUX0P(DfV&-0MRYXICLATEHty7Oah zm2+8+DS6jjid&u`WkPzpiVd7xcUe94NV(a!bP6#$P^n&F~*x!9fhhJ=cunch^vw46A_frOjs(VIvx>Ed^@^NBJ z{QB#J$f(Ke20R~NY(Ebb+!*9je2m~Y(vl)j?_V||S6;k%pHslkNhD$D0~gz*EE)`^ z8R)l?TjwB((jiM%U+4HdFi~<;ZYfY1yaa^ZjYrbmbR_8~bk`z0w0LT~K4-_|avd9z zrKHBp8(|9Xz z)L8CShXmXuRimkqQq~$5X4!}@P0EWzLnoK=a7#Yk*%<22YqNe68V`N-6La-OjYZ^i z7(E_EFwyf7D3e4?{dxcz=mrCSAn9M?c*J@#M7Ls~98PeX8@5;HnjG#plaY&FmQLxy z{0!6y1kyTW7vEmL7{~F!KS7K3hxDvtl3=^HCjFpMl@{H2y}x|6|C_GStdSBDDBfF` z4*@e5&@;61Eo!XG_ocq@!x03N?b+)}g|#1it(zRMY;1sN@Vk63SP&%S*#i{x6Ka9X zKn>e@aLF(##(qIOg^!&1Qc<9{>GhwjfE)7JHrF_C=lo_lG0{HV=JmCIv8?fTnA zlX@EPC?6wze?1VW5+8N)8G6189{c*P&k<=9AK!r?U1LfXV#IcV+js>rpq@VarVDFQ zwt&e;1~PTWgVLWi{FrOZ6{4=lv9Wz*8BC{zBgZ8h8<|cI4|q>yod8lxd^eQ0v4y^u zmS3GVnD$zeNI4{kNZ@fi*9t6=7~4Vg63oCR(bl;}u?bI&-!`A%1p5=j`z8|M$l5En%p4LAteBnJg&TtD4ja(&f`Rd`zAHLv!<}_R{uFX5NhLV5Eeu<_ z&`0bOFeQ@z3~<|&aDv|U+c%vN2XfJ$%S(_$nC{ag$FY%Pyj}=wjGpcJ;osV6J z&+XSE=w7PZ*TxiVpZyxcU$-;AvG4cq)nk-~!AFusX`z)`s*0u{a%q)zh}vvP1kcB2 ztrf2MaZ^lE!=HUnFxBp)!M-~0&i^owX;v_h-s**JG!IsT{mxtzhe5Xma!HJkuI1_B zbn2&gOkO-c?1Ay2Y>m}`s1_K!V>ChkN z1&yhDLk?acYrXlCbMYq8hRqxMC$w;k07oxk4Q*I3pI=Xr1rmrPE(vO|gueqS?iZXD z&t{$AxuQZPWE~=~!R^d45+mos+Rhv*Y77?!JwZK>sdsV%l}7@9J3U^A7s8r?N$`&d zWJcn2kll&tj4VyCR#tC%z$F`n!S4rFvBhLvUj{PUpKP2&z$+Dcg)u=O+}~gcxd^x5 zHTGm|u0j2v(+RAe;+yA9rpp5u5~iZ|?B(Tt2b zTMBHt_9??As0Q2-uz1dT=r-CD6}VG!do55rrh%!gyo>l>P^nq%YEEG{ZMO0N%HS4! z3|PzNnbdGl^*phJnB^U3knTR`i>Uw{X^xrtom6(~7S{yNronlTzCbYM1-tO{{C)QD zrgwMw-o5e{RfIA`eVPca`-nt8Jp-SF?Wd zmnU~SLDz8WzakHe?)bYtu#N7ZW$yIIg8nKzrM|l!GTSZGYItm{^E>CktGl?5k(myc z!0Y}oz~$}3wDyfnW}|pNFEAVUOeDWLoRoiO%!e*sneo(3EJhyqhJ7N=1<==T*7=1a z9lz@@SX0W(YC%3Fe>Zuvu~ai5;4R@7yb1XCnWAiocXpt*%K+^9LC%X_0Up@!5vgkA zdtCedL=MkGh4^(uZ(8+1g%l-72_c5A)x6!Du-n@4{aPz9DhijcKR!|t$VeAR&-Q}U zyM1b((c)+EhGHuyoa}tM!%+YSX{%f{D9wX*K<2UmUsVd2b_BS|(!m9Z*TLyF05DJr zHIxsH&Vti~6N@&LAlRlY*wxxkHq<`xvJ$m4;w3c+n4JS#ZT#|#mRxWG!6xvEY8pQJ zwWg&>G=HlYM1=`u?_njYc{v|B-H^R}7&!G^Ekm{p;MDo4 z9MCbdU7@CrI+$2}WRNH#Xj4Haa!IahKOv@Lg-(tu+B;^m)YB^?mIbluezn&*>&e2U z%Zp+`kiFM5Y(DvTpQj2PQVEzo1W#*%DUE?Dk=(;J zr8*$s9R{N$*N{f6Bd>s?%8v-*A#MK4|6Z!dzhc|qQFWi^^^z;63%7QH zh3NAle$od6V^DY}MH^{JI+dC;S^ysGdnf=Ahe#94y#PtHMBq|B_=l~=lHM-^WnoOCrd|9o>~&CeQ8)qqgRBOQZxj9HEq7uK;d|*1 zQE&}dgmA_SCO9wqR@2*C-zsOkL@ypX8o#&po8ttYUKY!Deh{io9mnpv|DcRIWv+0O zh+BSLFQyHK1GFDiKn#6+$)JJS)e^9I%oUg&`yIj2cfv>>RZZJ)py}Qx*I{*qTTcLN zEZFM#^KhR7h?3Lq`B0E@A4gs!J%8b4d}8iZxH7vD8apZc2>EgqXUSav)a)K3OOmkruejxD9*0?Ua$UzX3mL{7d{aZdfW=u& z#>5e*32zd96>2A~%_<6dTWLcADL`@JFCTZ+qgZU=g{aH~ScB0YBLXI5o(bz)6ndqh zyCBY>r)|`C^Ysd-tSf7x4GSH+I|zcoS#pzz$>a4A5|Z3ZV)}#wih3rkA^@SKhsU`P z(hzawnc0bpW3R~8>%M(9doKMVFK+1}%oBOF4psiKX9707;BJ;G?-`>GdGZ((dG!S< zgFZbZ*Ep`DH_~YD8VIJMW#l5_!2emGj*X>>Hl~;j@;^0^s3&aq_0y5^fvkoSG$ zuOR^VmfG5ycJ?AX0KZs~@+j!#Z%2b)KIFH)_;o*$>sR(pXZ&8_nFBVAFN~_gQ)9-=ck}~W;lrG4&ge_W z+Z1GIlRUF=iCMKlm&|e;z<2@}l7X;2EVD85iV!J36Ej?@?&CJ<=xaD>)xs#;ih%3R}!S(V-k_QPF_e+Bt9s3o#*Wc#ZC>F?Y00zzV zHhN@0F{wcEYJi{{z1B5-F_*p76ZI2*4>1%r6XZ}k{x6%K9lpYy6G9Pe(TjpL{ArP- zr*gE_Cp&(07Gy50^Vc~NzovFg)`q(!4<^qascqvlI;^p+bN7L$L50h67RFK@nb5D5 z+3PvD!b=waG1L5FrqPwr7{4fH)-on|^AG3!rZT1z{-=pfZq65|4)20EvE} z{B#WsM`eFk7anoCE3&{E9C-+EsCeZmx|~p4egnQN|JWPZ8K)TbpDent^FIHiK<8Ig zElObmUsyvbkFdmQ=ORWh{H+DAxAueRcD`B)$~j;ZjG|+&3<6me`mcUxMI?B`Y6cGP zz5|@!;WuH|EN=oC12F}c7)PM??69YC(-Iu7*DkZ;tS1D!9W8B5)27+#H;WE#y_Vp{ ze(V9u4!79%U|&jv8>zNo?m(89zXnlCfm*A?`qquB5`aU0g3Q7%7&~rn_e*T)QNQF^ z&kOCx(i3Uv$D@^b8$(MaUB83R8sHKoF~CZ}XN#u64+yBn)+%t8mNR%&OTlXOMB~S) z8d9bG&kCS2-doP7xaftLnUl0o$SGneIn`k(#q^1IkMsmXTa%b)3qdba7?hI#Xk|^F zyhE&X4j~&Oiu5IPy@Ds+?tbwg-t!Hcq!sTD@J@97!wpz3^__d;yoYm271FqHTrIiV zZ{$e?6x3+sAFf*QXmCPyC^<`I6zZ4&Z~z)^y7qaOR_ir-x~E@Tr{nao@&)BPcMJjJ z&cZvG^(%|aS~}#A_C1jw87+A2c`_RLWOwwG&Em3r&AvvGWDI6Jrxfq`5WV-dl;Lep zBfFOGP=lOiM8s$)erO^)oAjx>eXiAH`GJ)tn+PJygrXzP;4q`!nfb|Dz6g463jVx@ zOs)5wn}g*QE-5^uLF4YXOBt1`NZsk2YKp&Gvh-cN;71I+){wmRy0_DkIJqY1 zijc_+5ETz)thzYf^m=M|vpE7(VFp%JWdl|MEWC^@$UnT9jL-|Mu8)J_t&?>Z_=Lhs z&pFs4ES&)8ikZ7ov`rTZG<6}Ah-Q!&^e~^2ml>l>QVtsul_-uAw53KCqD;9$W zHZ7bZIk)KKcwy70QdK&0N0LH5Lrgk3UWOIr5~oLFx#(ybXs!UPp9Rpvcdx%?To{s} z!!$r3(cr2NuX<6`@b+d>Gs*4~qFmM9_)riV#*Z&h0b6z2BKIhvknph`p6~0vE=X_Uww7PHa+P5H zDSGT>P?6n^4Pk_AIN3%Y_G~G?UBJhE{eI*yaVxwLNWLPG%0Ck{_e47)-|39 z^DtQofq*===m~AYW@Il`gdWGEGF!pIo`{1o0zCt^7G|c{F@dQ5qj<;U^zXY47-We%AI)Sw*0PwYS7Nac|5LxInS?A(+UtBnjPyv(UL$uUTU- z7FJf%hr79V(Xl??YoBk9Tm5JzNg&-roMVY^ z!P6Np&OM*rhsmJ`WPX*m1`Gk19%VPQUi<+)J3(p1gVGN2vvdRr4XLoAoD$c) zn!F%40KDaO%b?}Z#Bw7L0#oVwCueC>cy`mT+RwY-Bp-EGy*c>6If8wuqC#BQmk?Hk z>`^!QNGN>Ae{Z`2qWyx}M9Y}#B)h4z!U)5J-r;`uSorWZAylNCjjMye_`*YD66t6`MIsO@SwCOf4a`E?fcWTT zD?R5=^6)BsDr!0X-D^mt#ydGf1gNazSI-CPJ!?PWXC3xI?la z1-+?+YHYO7u8LEbobw1g)o%dm=+9RQKm5#RFN#M^Y!i5a_$-%olZ>iGml6LTHQ8S= zPAuYS{{vK~K4fZ3lK7-)i-K6LQVPBHkbUSsxXRVP5{!IR2cIhj44YwM5sTk=&h3kj z&Wqz#TfzFUNW)GLc1f^|Nx(^b`r$BQkYZx+<(`t5(4)5{DGx)?(_4CYA`NC+ZRwVI z_4JV6D>LT!=bKEO!i~m_5t<+av3F6aSw$%#G^qnuIL1GO7MwPNEiUBczI%_`Tl#`k z#KCdk9Y6pyBq^$aWcs1x23`K}NpjWuA$#e7WGF-Wn#q2~*V2cwFObi65?^|e@)kT?w{2mY5Y--T(5w4>W0d0FbU|=Jyhk2Ov;4a_qpOdxXB#FPO(sP zs+$F8&<9N@=ko<+yP}pi1)hk=9HAKxiuFld059V61(WAiUtKw`aj3~e#PFoo!*SSn zHJ1&EJx4s30Fb%J2UbR|30^`EJw`nAmS0`ARDsU--5}88+RaDr@S3QNM~SgkEPz&| z{1}YQ4B1)+|A6NGOiN_&Av`GsTRd97Q<_ zV~_iUFpE1y`Xp}s^JZwi*Ip(A&R`;Ws$oV3H-MGQ1W$~W2-g;aa83LtgxUh=)s!*r0czW6GerE$Ot2|;H zM2!8>g#c(+o1ZQ)`P6<~S%l}}xCAvuuNVU;oRN94gdNaxW|0xVs?;fEXGre z^L_&Vd6%?tgR3jtQPEa`xk65pN;a!H*|+xrNP8ijL&SQ=JW;v+uT;@i^gI`)OVp!#m9%$Y}rkLn~F0 z+KCWKH~`0w@|Yt#Hc@3Z+SYTAx|zPO`K_iv!JKNkm8yMZdA3?5c>y3Q(4T6z-@XZS zT>a2aZiLfYekf43;_rCO85&|C2q14dck*9HAH>EFG6aO0t4bMTaga>_0feEU1b3&o zY8d<=h~k9-;bFJGYd8$>clKm=#xNRfj3+d(=}{}HnNdBc+Mr-lNkl95l~@QTTp+?SFzU2 ziPp`}1#vkswnkG#_c$ta&W!?qnrRvHi+vQx65 zyBwWukCyjO8mT#Gie5cpyO~$+c(QOCavfxTuIXyg-n;CITUM$%{KVyyAK3j>Z7`V9_?*ERt3%*&AH3 z@F+Zmv>=b{N}rkiL@184nq}o!x$4JLcueqq^;|!M%k;V5bUzJK6a5qr1l{#y*V8Y- znkL-7*TZF8F`Xz0nhvl%@DqT22)H781Tb+?P1&&h>qV}u41YKHuYflz3PoEBg;+~N z;=ksQV>hA=e0SEesuZz_+;X}i1lB?v@9RWE(r|qT(SSVLZB^>X)`M2z&x5*M1#Lz+ z1mB{^BNo?ZS>!==Pbcmy^Gi!*Jo;MuDQNgms-d^~_1RkVbbf5h1%u5ikM~7bw5NvbVd#NTphzo*nj@VrS|*@Xh;+*`D+k~8 z0cE+HB#Ctw#j}MDKX%$h!gk#l(!s{xI#z)1b~$$2>9mKeO}(9Ak`~yB7)9W7TGyWn z9y}IVm_N_aC%ZO?bTd$J2A8(Cp{4^B6G=2@;Rr=Iw+1UQ6b?nk4$yw-<#e)Bxqlb+ z;VS9SAw90@H}s|DjNs%d&E*D@vFU)mm|QOO4U{jW&hh|x)Hmjsu`}YMBjSDcPELI- zH|>c*d7&b~8?b9QSs_!}cS-n0i;+eSofsNT56(v!It!E!oOeEQ!2J!EeZs#%4*;)^ z1g8z6F=L>GKqim`WpjZb#~WcjhR(?g5Oy)8RB>z65g|P8hx_;NIz5rhc0gffAcUdm z0(5&HxK+3`X4h_SBds3n69SK84==Rb>-pVZn3Qn^`bG7MzFjER?O0^f8NnHqh?E@L zI=Y`*RpD#^>3okI4D{`tYs=k^$~!wLcPSJPY^bV8%Nx_zSJnaxVYoju9}d6FC_#X4 z(%^Eo+xc)hLuF$|_OPK3;LuVj`D4S&poMj4}b zW7Fq*oplNxBFCmCKX>79#fcrdSSbXjLVN5F;$){(0d}wQBM`RX-d>6j-b@U7U7p0| zKla{h=~Zpb5`3DB=)R~7j6l+*5FTL!67Hr0BVmL$67cjjSbLv3nSH7%I-;WQa>v^L zS~5R&GUsQ8S?Y1m=Iv9uZ9j8Q|M?TC+oGHkNO20z^HmYN{bbxS+KA99lJ2b~C|}`$ zhf7UFKs|Ag)t6t%=?P(PT0WM_O1IY1o+P+OP_GzO@CqJ-0MtsSp4`J@_mj)o$e^mP za?FPXKW3)zwg#s3g%z(r!1P6bE%dH+4U)-#8dMb=neVXcCL|U{nfd#KBCrcU&=7Mntdvio8BUTc z-hOMyN<^?@v=SPHDQ)Qs;s3Zort=!LxZ2x0KvkV;LRz>hAoZQqCC9FeDTK8N!1mVr#ukuoaEQ=L3obxl?@% z5bv8mmfL2wz&WzH+=rWxj8Lx(En+x$U3<5nWTP)MsG}Z%2;c1Wg32DzBLsp#i;En=-po{c zr{CwH_S!ju0^|graa;l z|NJ|ZFtDV;^B)99IBK~hLBT$GCvkim#9 zIXeJ8b{4uFucpPtT^yIK(Xn+74mAxb3x1XIpr02JN28}|FL;~UpxZ&Ke&sBJr;Z$ z?CA!puZ=6ABaey=`D=I4IjV;)#C-nBWw!^@6K z>1o(QqM>+e6?^ku)<`9f0-3i}Zz!f9>9+4uONo!Ta;Es>y>*zu3s+v`9EcLZ3p9wc zom(qeIm|`yZ03GgINNKR)}csc@o*)x+uCXO2wLs$$M9vAVY?*2w7qnPGE-2PJdubUU%ZB4g+fy81sIKVLBMAn^ z^@4t1&X6ZVzbY~`0^I~nYPenMKWCHoEPHyhx1Ci9a-CeL`Zf8ic~C;p`g7S-OX2rb zBh^L8JXi3kD9;L>)FKKY@dgaToL)f+W;{PENoeB2Ul zCtA3IEnz5ClmWhJ*~O@p?(T9KiDw;!0>Yx$etNlpkR=k88lN zM&Zd1^JY1etUB%o&5dXqb%o){Z<&E~#5(tMDTIjMiw1WQ4g6%yZZOXDFYlUvdMCWi zi?0+#XSZ6U?BOV^__UJsvp35JQFuIkohdZ4{DS745R>L|a8fF&BM?5P=RSDgs3;Bh zUIK_o;ijkmCKnXTIeAze3kY>O>9{9I;Z%1vBGb52h$J}qou{&_dTHH$e!a+5(4zaG z!UKO|dLP>SXGjEJR`j6&ratGIzdv8_B~S~9NfW9@0V00( zPuhM`jtQ;W%fV4}#=6Rj7CXK(YbJ9wJS?^$67BohZz<~{*doNI9Z?WOR+hShxO9oI+kw`soUr(f|@`JjD49w3&DZ8p{kcGb9=Ar1RrGG9g<{^5q?i zw}M`bMSLa>asMsYWMjHqoWecQmCbd?FnUiduzgOcl0->#R4krKlgvU!ftIJ2ND< zBjXM_Z9Z^07tm|y&1{JX)a0OJ zR><2%5oK=}DlK*X_`roUxrG!-$D(;-A;$tnKaMH^AM9HY z8F=Y3SATTy{()(i2m8+Zv~j21Q8rg$`>|S(yyVKrCbb>Ja?cUq>1&sT!gyTkBIe#x zu^!&>xAH9r7%UZ>WQqZc5;hiz?j0~x}bi@V{Z zlea*269yAaQs2kSzkwJP0r`!NnS2iml>>BI)IaPCn*s_Qb-Kf71C8o!;XJ^ZBR&us zD%twLD49>?1zbf+KvFx4h0F@d(up}c{_$`wBXfpH4+`Eehv-)hKMD`Po*`AFVg z+e2ko2OtXoAg*~q?o=m!pmRnKYh4xa*+@wN-~^b03-0_65DGCwoFeAy55TWW3*Q|) zB7pc^sP^!ip-R_+N;z0Rwah*&CJiX<+l03}ANL!4)%QV-Qml9AbU;J!kQ_$n4x%2S zaV`JqA}qUK%smhpSZWRZ?-qp^%nF$biPwPfuXmXMB;TAh0fQ8Linf37-lx&h`_WC= z)7g56_IheN;z@j1HAfNm#t7TV)_tRfCfCz-+&u|Cx&{sDBKD3L(0k=zz36sm?#t!n zsu_0tn*fY3M)VO+H%u4-un_ukgks}Luk60Uq+PuJc4G`qz5PM!t5+y}!ieb3v{%S- zE$Th4#<;9Mbw$Nd2td+9;PlG^#4|${G8I9|6Wa*{3X0iVDJ9P$&C%ICc-Kob)(u0` zy#Gd2WS}vBlfd4>f^_3-gv!`yGkh}<2`7_B#R#Uk);no>lCNT#8GxVJk{dRV=QIq! zh2L3n7rq+ExsszmCqUs7odik&!-11H(tjbf88Uja2jNf$Nb!Chq#v)} z?wjXjuZj-T04KXJ*3EjzrXvxcTnb*0Y+`N#W<-WI>`%q zBSx4y$3*B*(` z$t+!SO3z6$A<(d|E=+pX-y2Sqy{;qSgkfNC>qno@_LI@R&(r80RLJYX-~2u0uZ274p)dVd!B?B@~V zDvVX9>Wdbl2!w``pm<*l&J%48X!z5dG`sYFay$(Avq&m0D|tWxAp^J78L%%Rir0nW zRTn6pwX%NMGT}Oun1sb)5011Qt__CW)eqOBY@hK~Gpw-o=vyiHp4wNVemB&W3y^FH zTR@sCi8AW)`28(hXisXRJuPbLgjwVa^7i-urF4_Z)OqAU+-}AA7*%Iqm3TdEQAyDV zp_ntJeH%M~1odL^3Z`E$8rKo}1kqGmBsPeN@pj^@qmB@YgSge5CxRL+%+&aXQ;|L= zKp5Em>90K!ZqjlmTK(>Bi()Z9Yyj?Z>C)NxhMe)b*eP_o=s;^q$K0Gzd0wt?{pD7u zh&L=$#d_yO@R3)G!DqPUm)`TPp2}_DqsYT|z{+qSR1!cq?bx8a=e+)KmS(sjq9{LW z=eBb|v{^6R?Z`p<6ut$04>2=kNs9SQj|1`glM_%^H=CBI9qlX-4D|nQDNZZra+2y+ z!|sa0g<$(|oe$TX(zEm(or{o*$5$X& zx6i}H10($|7z<+Fqb2kfxGN0l6)k4$J4Gb#Bmd*$0KHk&u>AmE1H}JP-P$i`BwwvX zADgqoQ!iX|^GFjSb^0a#bHf7wA3^dX{zRQ;vKqrrG<;LXqa;#2P{NKoK@EdApEUz| zIFRBOE`%vOz-q&3;bepFtt?Z!+{=deo%18M;kJsw3wOD+iTw`mN6pi8T?@K#H5df@ z430mSKeovi^YHFpC}%zv^CqApHeB{FQ8BIVI74!Rtu(Dns*^7-kqt{4W@b@#7sG#} zVVh8NRy5`RM)O58zNY`{Nc}sFq9>+kg2c7o|0e$Zge$iHi^Z5_{%`!+zft%O@r9l7 zYr(JLCBD1A)kOo+=)c~g`>Z*Ch)(2R1TdSGT>iJ=zy0|hE&LmS78a(c|Medlrd86b ztzXly;(Pzt&H3*gt$*+KME*Ng_I=mCBK>n7c*P$7ouvPIH-8mb{`Uc~^uO=G@BPcs z_}A}e|6>pPyZc4`-ya)|?vYq8*H=~jFTeTcJ3|ZP_P=iH@86e(sDqvm6vqD3cm8|| z-Q5gzy8qV)|9F=oiuA8T|9@7I>>1xdh4K_lo*!8|IetHl2Kd3hz7x$^jr$QzMcW{E zrvB4L|GM~puJb<+haxi6VE(?ie{HUdvcEU?_qE@@;S}3H-o$0w_nOT`3Y^)091T<3 zR#Vaa=QoqKXafBD_N zug`ycto~0nXUB4?{Nx4_oBm5i?f|0ef5Vf#-@k8E!7A;~?fdJEaa#Xzv2l&M{j;L} zGZwi0$uDp(Zcz+ht#&N-0fx^Z@;S8_x?Ht?z(vU?8``-^uWi*1_&j~RqZznwH%1CqbqCI0cwzNu)6{*6WxYCj^RFWn1k#Qt=i z_t**zlHpZ&jY{dc+8SNQH-%nN^Ul*=WDWz8b&YsTq*k@4{b{qPI=f%qRU{_&r1 zz5Kvt&l0Gvw)=+g?a9WcnZRFuiU9n8_}66;K^#$VeeO1AD_|`Nl0b@Z)KeFc*SHVt z!WBYWeh^t9-j|O~sc3ovP?slxWVe-CBv`as!gvHU@&@)!*ic1D6|gxTuIPF$nTK9& zfbA7FAnL&@&L-0)h2>!nfn`t1(b*_E4X#n1S20Cc_ZnPZpv%k_XrXA^bn~bQ=-wH( zY0-UYIhE=<-5x;bdti`+6l$v|SaFYvncpq@0F6Ryqg&o%n{ zTK(^@6TXem5?|%HOY+xsx~uf%f~&d+0*A5*m#fIX(mb*CLT?2{6g0byWe!Tb+w*&I zgXU&m^WbBc0_4H}Tr2oVWgSFMDF`Hkqi8@vJ(L{*NFvg9-71B%V^@ZfTZdHMMU6p# z`c9+V!Dg)~cgzQE7_W-h{Gi1gato(+887;6X9pYW6o;!ji=Dsy7-j zSx+*JeadK}!we$xD{jCl=>O#^om&hr9Z!s6;~Xy*B*Di4{zq5yrIEWI8K7k_`{$j>zY{g}nIG0mtqerJfsd|g!g z??c6kBQ7BXNl|g&bT@{=aj(rkO#Ki>d8dJ%Gx{k z{P)3!$<)}Q;}Mb+b*jQ$ENw^O_qTh>F|1k|EF7-Xy)N!I<1%+v93=hjy@@>Im_~#?a|!&5jQ8 z_T`b%Fsf+LOo7pau}&d$2fowFyQcf!}D|^6Ua|$G%SBA&sT=76artyHEwRb z#h`T$-T80Li6Qa_Q4;Rh-&r<>K!C}G@LgT8+zbd^H9u$ahI{Ca1cm2Eec6XvjEwvz zE-suxC$#)tOMILZB8%a+x?ksSrf66Lk@eq7Bd+$+${F=^Ka|oquz6q$`&A${o){|A z9)DltibYTo!{=<0_n&i=tcwYTyhqY^uqygfZ?{dq;gt~y{&XAd$1F@c)Qw{d(`X_Ayl|cp7emQ(OBW+&wx&$ zkXro`oBS}F4gxO>-+TW-8-UZV1?p_BJK>9SF?h*!WQX$s;y$UzsWW_{Wy;+sB9Lo8DpWWSSo?{r7fwXG3I&8%F8h?xE4qU~ir?%)pW1KI(k`%)W^x~)K-UTc1MpkI}QchXJ&n21!b z)c?G09dE%-BR4cNdiXot_n9U*tSv-IpU)12@COZT~f>p#r??{f=C zWxiF`#(I9IANZdJc>Dc^kVbHuB^(lVeqfrxBgmJH#Y3sbv=qoy z&q#YhYhg%9R(N>}=&T;x;yOEB``C!ou@PZW9Y~jmq4w(=b3htn+tz3Lkw}WV* z=#HJ4`~ya75a$e-M##VWZEBgJi07L@!LWl4xy;1CHD^G8%Hx0w=He+iIp`2-^w&=! zU!8L?NtN%W*(^zV38rj7RR!ok~}h?Cc3>%pb=nA^ax6VmP+^ z2|HslagO<3s(M0{aOP@6B(4lt@V#;MJSwHEE6cAvmsVW6?`Jhsq;~P@FvTB3-6POiC5z?8#mu zVl3m)>!6javjUqFx4+t%Q|jj!l{-kN_{=NpsiSJ34rd)6I43kM(zv8;!$^B+gqj66+L;W7=H?tdbZP?cJ#FZgLJFF5LPGH83aVM=FPQ^cFZgXftParjD?7_pCGZz%73~+jZOoGL`R+ z+ll)`_7qdG#e@V`gP}qBhC|$hR~`35HLn>kV5YqR{Eh0%=~~?qX@*DnW=LXx0-2x~ zJu+D=n5m}Xt4!TS>I z`A~ghi?;{32B8QhYeFjUdzUo=&=*rnNf*AkQptGDuD&4+*QtM|m%7YzC@HLu*@DaU z)IYp>xGx)@qKG5zb60L8(9q^Rl3h3!8cCCHektNP&C^*S5}_1r?R(->*{9@lbXuX_ z!3oerpqu5r%12mF;VNJ5cS%b3YQR{*vRE?CITgeqaXS zV)51Ge-@!n^_A~ni~usrg7Swf!`DKm3s4Pp{!~?N`;~rHk}+PY!J9JOt=0ri=Oke5 zScZsdjId;Ktj*3kwY@RE7qonNdLU$+pjc1kSp�j?p0&2QR(#;g2>7Vw(mB^wGm` znbG@0i#!Ha3pBRlVjwNHld(~=~UmU zhi_rlZN{le2(nL(Hq0##bx&=#gVLwhRS=kuh#;L7AKVk&|^Q{Z%oqmTkQP| zk%zl^aD-E(41BvA@ObOT`y?KtcQm}8?cs={fN3Kj-oPIpJIj5X`H%bCgDAzg%nea` zkORi6>B(fUlTt;*pOB4%$(u{~xt*GLxvy5wDD7`f>+h!)BKTi_d;HY!e79=n94;m= zqDTe;LgM>qtIYXvs`Pw=zJNt*%ZpJ9=Tir71lGV!4Aw_J5_n;f4@H8<3Sl%SFg9St z#w@J+bC>01ouS6?C1q+~Ix%X81BPh7bwje{GmIZ+u$1=6FhpM^YC3Q0fV*S%Axcce z>9zyI{Om~v=mVA+3BmMZwMMCuZHPS2j;~{4exz%ANR~W{e&_ot?h}x<(<42uKZTxb zeV-~@InJwVyrJyiyklUBwC-!{4KZroAm=0Ee&ydqX1E8i^=5rPwOXCLRbRcijfLbD zU8GJK&Bs&t;WS?EElvzSRNwPD=Wdfw$uSo7eWcYExVIz9qjqw!P zW>0CE4T6dxQ}k=LUgJrIo<4LzDVFb-2Wij0hxViUyNkwry*^&iAjq9(bf@hWPcm^m zU0!rFoaB|-<_F2T)cFP^Q^Z_b`hly;UrRAO&sXP9>yThSpOgEbz>EG=2l$DQ%n=TS zulj4p12=6*kZ@+Cog%OlGq>T``YHDo#NRQiL4@ZCl?GCGzYJlIVCfdf3`yjYsc zKNB6CfW$@ngcC02q%|sSC|?haPiq76{QQIeB=vX>sBL_v$JHVAGq(7BuK4NWfBXCE z(`qG+uU%76`0ATrW|sQ;$F13$JoslKhR`mzkxa440o6~>|9**R3qyF zR5sUhvXXUzz>!n_%w)&|-@@1MO7WNQb=$i8>H~eKxgU?V0h3gCiqSO1J2KS>FvApJLszr^lV& zzYix&XCR|(zQrylkSXbwytUPjriReoMOj1;Pw2t8(MKsRqwvcnY|ZUuIfK!p&s9T3 zM5Ka4+QEBTdu+EMp&y3&JnBTwd(32y^C3nk0#KQ2kK1H(4M@5!U^+KEm~m;Zq?+F5 zX+!>3Hyf|to=&JYJ%szAFIf)>bO%`ycRul7@UT86r<;eFwy%elr61legI+}a84QB? z((ePZpyn3B4fA~XG4I#y_EuU2`SQ2rRArB|Z4!n}1O_2oyUF!kG4vX78#uwJ1tMN&-NrXufpvLe|(3k-_L>puY-*}jQDNLD8@STGW&D(i} zi^X1H%7X>#fUrd6&2-F{1LJ9oy^T4u>BX$T zt$%;tpcD>UG2!vAY<_#>A2J43@c^21m6M( zc*mr}Tb!p=ckQ;gAqNSS6XK3sV4VOuJqO~0Tns|S?wjjx*U`YGsZ?$->&cIK zevGK>7e*oOjsl|d-OFnrOx#cp7ZgC>Ek#V@;3kG>G}B}1*NnL=d%8JP&919}^SEt@ z+c{NWejviyqkm!nYz@!6yB@{^`QC&dfTX=?R2kZDr|;8E&9*K`RnH6Q_j%>u2+uY% z&F%5y77su0f&*Jm<}yQoFQs}MN`Et7-<#bw!8wZ@Q`RzO^QsKqEp&|i4J7{v%lI+lRMq52>P;%4wOk)Aw<@yo$A?*NY0BfM(EXRkN7oJa4o91{Xw z&uRtw-qsk73XNtq4Ohe+lk+0W+Ps>R7vCJ>@hOYuXs)-8wr?TEY9bBwsj1s%3Sho; z)j$Yev^xMldGPSRzB&|)?tT~*OiMAB&olW}8R_^(B%$D+q^B=-Gd-J9xs*h}aER?8Y#fTQI-mS=$c&@6uU@ z&!{P3V`dHYW#pS0LW1$Vr*faXzSr@Sva9#?-D~rnj$|(>bW+oU=O$F=fuwf$=|$Vw zaq!twl_M3_DfdCm4f?J}JJ`<+%ITknw_a0>r<}47Epq~C!KY@05%chRf+?dd)+?tz zu))-P_xzucS*bd0j{^*O%E2k>Iy^i4Rte7D)R}&l+U>KG;?dkO zb8pe6x7SbN-ad1*86rry&wMCO;+TdHr))CxBVO8kH^w4h6dijg)N6e(z`uABh}(60+%C zHD+~`42mI+MkFviZ)(gwen)uIH@tl5`xYKLwA*&Hmv_>Gl?bXkhR%Ld+l~Ff$8B(3 zv+}`CHnWld}B+IvHHkB}8-1TDSyih|C>6xFvkhYY} zDk0Nx=>Z~CTHckG?m{N4U3(W=v>jirri=jLP;`eLasp12!pEbIPS1`U)EU;52>+TZ zT1H5@%<20H4edM?WGF@VUz||(=H>C{JXf|ejjCO(+UPBAzATR~F6MYO#!y!JJnip` zIzT(wH1tpm#iOoX(njqt*Mj?E9X~sMn>)R-!5}a1FK?o@e~v~E&tS}74rey*0m#)g zHiC1EE9&jy*bY;~V5tQ$5*IKlzT8cX9heG3vi)i0)AO??rl7IZ7EA#SXce`$Ja&I0c#*sK$*B*sy zpPJB)xUORUJ!_l7xaAwVOt0#(yL3C3Rc6eUTR6TS6367aIa~CS`__m zoaHk>yBA4=e{K2ffB*p4XZTa2V19!6078Hn(P2jrEcrW{GBx~}>#Z3+t`ywxKJ{LQ zA+1YqhGR6MA|Vlb=*6mg<;c*ehC2K%BjB(*B0H9DNMK~OllTsb>7O&_Fmz&E>$;&j zb(O7db7%fxm=egn`9||@AZWFfmdzK|a*c$q@8FOQ79-zO8eL|}7zuJUZevsF=$VX1 zmBZ%WRq2S`zP_UC6aBr6F0Y4E!s{YdnTPJJleFIdD3sFuWJf9jl-T*9vU21D$yg+N zL%@?wpVE)MG#r3akJ>cHq3!L7v&fIZV@j~-uIe%PIUH-NZJo=M#Q(RsfU*Fq@I@s&4ycd zl>_o2*l5FwEdR-sTk75SC@Fb+{ZYFv6tj4xLMfLPA>j3T1*Ifi%tfB8;eibI1Cnz& z+{K%G=}wVwBrIPv4+S!X_tvg&fQ{fAnqAzk5Vm7;Vo}%YaDrx~bVw`i*SIhdci{C+ z&KQd!s0vn4I{(aq;5k~gH$*`sG?wA!S6jm3^C@dhC}nOC+4_5*mm>`2&Lz|+5VnbH zX>|_=vdiBG*%4vbNc6`w2BD(J55x<;3_Rj~pC+6c)-_^7G7hQENYJwQgyjs?!o_=| z>p*pHtsgH>2+8!~(K2-aVwaKl^y)ToLt@N>-GVl&tuX_}YV{%~tc%=Jn;-0TZi%BU z%xVi_7G%z(y?@z5rrE@;(qzHW;#B&)&Buq2>;~|D4wg`?YB1?aasdCq^9D$Kr6 z2ZWtq3(e%xSa+%WihyOOwJjYGX}C==J+!QMgmLvC`Lxz00w6 zUx$+Z#E$Dnj?TykUAXVpEm-Hsqm~rebfQz}Dtq8;LtM&iAi_)tWY^lTIyl4e#(;3d z8ZBu^Ts`^{J}qS<2Ga6UbVkURcE|*l+fLLg&5v!8!t}f~VO8)DllJ1G9<{#624$+@ zUiRfOO&01E>0We!amtTaa%C`iXefwX{BRmbgu$^kevvty8!ZZzHxWeU_DDl4^=y68 zQ~%|9ZZck~1-*PDWsGs`J`_Cku`DH3ToyFdh88r=7IKNxWXm3XNHv;vY;nS-HJ7jM#e)6{m?PfPJ(OIdM~VS_UOU4p zHrXeE4GlZw0T2-FE56@+;tW*QBnp`8#m^)(pj|8G+t1F6Z&S&YjYTneaGmNw!A<#f z^7&X!4e&6$cY-nKtG^X)@X+t8nMrATWY`0?)A+4as!>$NgPgn`yg=Q{wMmGYJ1slP zn}HM@jME3X$ijPSWRx=NUSY^w16u2<~0j8=u^J`?NVwC;Gu*)qFLQ<@93N zx??<(^$s)>trF(y3fwQtJ@wIaeqY!i)9$^W#|Cx>8U-#Za4&^@H`=2?Sw5(pIx3yE zsPO$LlkcUtXnbQl zY{t3EGWvp(5)7nB2`vrxLa$J-p`@%20PoaYJV|{1h*$3Am#PV}1S=5j?SmuMhf81{ ziJ=2d)1&zQa*YN{@1F15X>EvJA-(3<9;e%#t%jMH0A*b-Gpse^a`3Mh*x{tau)5J4 zS`pDju2dbyx+slDk*1RR8;j@z)+2Nn*GK67P69mSg7*ukhIw!#Pt8P_gGWU~J8+x1 z5tcWY2j=%x#2AJ@#I5$9KYHTCQfZs%!1Ila-sG((-2;yrJt|N_n=cLUbDt2F--r*S ziJMthIOnK^dQ)$UPx-PvIt5m~I9oDBJImgLs+*X_eNs6e=f3Uyst=f{me(gA{t_(2 z00{s)EvEWYO5B<>qPKJg!*|Rq-v_bD5QP8?HNFfWzPNKl`D8#RPedkfs7s!Sbc5LG zvl`QDG^LHtx_fV5zLj8qJ1j)|4@syn*4O7A!2PUujm8@7CPlJ>1$O)OhfAj#){V|R z+w;&{wx8;rd3R|QOjKqcLkP@mOXhym)Aby87(Jis$6m$au|Bet zW1o(X>HcdRfxBda;M8d#w+(HzixNq9!7n+?a2Uryj?grbmD8HS%=@6%-sy30)!7Am4 zNq{&}YFY0tlXz%1<(_M~RyI zgRu4I^mv_=uMyp@aE??yr^hAB;tbX!c-Wk73Qlmi?ARCkTlcz17-zk2kMdd`Lw0|n z*9frQ`KI0n2VhUUQHSnfSyzUUEkq()vaIvEqUpHCPIrsn_4#T&I&XxWRahryIXR73 zyP@H>M>eQ@Wm~RRtV{v#Bz%N}UB%i&ifJ{7=a{WYx;<|Q zpW6QRx!s#x3;_$=g8DeH!@gb|P@s+*Xe+qMaS;Gp7E^gh2Qi9Tw^(7nSXpsB_J{0x zeMorxOB0OziGpQ#`vacz@zzgmnanz=6~}GbPWaaYo*3Aveo8raaO>E`73a>K5b9^G zuW1dU8_e=~j;Q)Q8BnFVBL*-?134#q1NqPxZ+Zr73*ZMpr!d`a!NB;l-0BLP2qrd- z3Pgv^y;Rs>0FXf|N*(@CrT#m^#jS)Pt=Tw(3t7W!t>b(FRyRs&=`XK=H=^$B9MkLf z+}v%IdAs($j#JZDFKJDe`-Bul_|co(Bx4h3=epT`^iSZt^e(!3?Md7(sIrwZ!n4WW zf18Av@2OOP+WR@?K@9JOu6+Ct6xWo&v2W>-t9YylHn2X%U{~M+^JqpL!tp#Zz3%T3 zm3JwQ9=kuXHi4Oo4^Yx(e+gZUMLI5VJb?%G z#*8JoPY9%3>G!FKpczKQk5rkz)9uEc4yorBv-4W^NTTxis*gI@9iLB+sg<%|RBn;Y z!^X<0vQKs%WonORxEE_53WAbwq|-@+H1YMc7{r390*>ccfyiz;_4|zieHo+mRC0f9 zK8@f_oUweG-&q?E$B#R}Owab>`AC6L(cbQ7c^+P;`D0AIaGS*43U`K+Q!FRv@v!f^ z)v5Nnr17HiP&`wJDoD^h#RvDII;m05z69Jl(u7fBf>o|G6YY`hz&Yz-z{T_!BbMd* z%?SMGE8K;r5xu}}O}PA#_H&!CG?y=$SUbVb_eQ~ckG`TN`@cc^F&*tLZV;|C1o*KhU`*Km0+Z+xm$MtMLQB9F#Z%p`8h2?|!Z`DU= z`%D+>E={+mTYl!c(hjP0*MYStOw(lS1+|p6Q(LnWs*!tbHlI87dl{~n|4yg;k%i3Q z&)$Bqi4f}B)ARX3`)Db7AON+uV2IwI1;4Ev+Rhyy>X~WU(5&$En~{Z^ z2iHuvc6^d2EGQ)&=uq63%8SEIOwc|*?eR(0Uv)4i4Y?Fz`SVEtLHzv&m78~)`gib} zggQT!-_}E_$D53`HL46$Q2}Jy=ljvy)dqy;wWMv_{E=(8e?UbIX-IA(2krZu0ke<- z@9O&^IR6Wuzw6;kghXyaxOD?>08Wg~J+jg{yf1nP%}(ppe~bYz=Z4MI8?+NyYABtG!|E%l31czyg~Cc0Fi^ZbmTR!UsOt4S zkPosaPK|%7mD4OA(>bb(+OrbDPnOGtTYz6<&&TCq4+7i#-u=;=dlW5xwck6U&!&ao#*Y48jgQNwu3=(&g?QyC>mh%*IOhJCmqvI@~~+Dz_dv*9j+8#`yIQ5x9a%3 zJ7!ba-&|grP25oqTq~w_`|97(_o|y%sNct>=5~#@;E1CtUFk%?TQ2opQnL5apmcMD zkmG0M_xY{1y|-~(PsL~OV-|gFUq|ah#InvE#6d3x_;!Z-qL*^ACAY8)P+tV8!Siai zAS{^FC+49n`P-{&ZOuvVy#`G&J^1qv)uIw?+2fu?cAjboDi*L=w86)%DCmilqZNIQ zzM=H2NsRH?f{pESTZ)7G%dVe~0apg*m}-D-v|VaMFCfomh;a8w{@9OD`!*~&Q;6&g z)j8c`D?58J0}(p`$0eI}(`KT*h?n2|8iY3> z+Z!9kU#QmyHOD_VFGphKLeT{6Y>vlXbL7I@H)Os5L zcZymTFfN!}@I}Yt;r9*DmbG;&3pW4KNy)(@!%TNr&NVmAv|adZ+TRQF(406u!sw9X zZo!5{bA@{do29JW06xeJVa4oCLv45m&KE%KeXYK(NQ~6W{S;m-K+ZLKWRX!i41C30 zLYT(in17$+CYP*Ry@SIo5t|EaQ}5@9?hGhPAoduy_as*qcx42XEPOa~ug=|=e-}ra zH~5$Rh_cN6M@bpsnGPiC>D*7v1rmDR$rs1^rkzeGo|-y`dRU(eHIeC7z1o{wIG}%> z`;39H;}htkut2a9EU5d~4+BbXN9EXS>CM{K`3V0pIc5pHS}y&JfTGMYx!1swb%^s( zRDKXAl&Gf32xo$%Ap|^6)Aj*Qb=xtj*N&mdJKa_ZWZs#nl#@ZufmfE>FVL)Ks4yBR zvV;6oWb}}c*-FQ-3J&(De9`|E&z2f%ik z0EYZ3NrwlwOV5dTsNe5lJY}Ih2eNkV!OX#Z4Gx7nb|fk@qFk6X%y(QO&adY6?SLuD z{G^NI#oPfVHj?zR*FaKaMH_z$4ibzg19*8yTRgpe0tKiyA2)$D{~@`7 ztOa5zcD12yGhUs3S`d`uH72~SON5kucgF2RcqcKRm5pXD08d%ZwAthb5R)$%4Fj0H zDSlmi^fNId*;2Gj9+sk}lVtpyKF{>`Z5CVOnMhk9T68%>1kBvS60aiZ1*m3Seq6`z zx?r@*!5)U`R1euXak|aL4C4#_dXyCQ-H`{(1I;kHNMX3P+63q>GvEl`Wu}yY)QJIy z!Dg?cc|^ZFwD=1&P@krx{I(bUeV!@uh?1pOw_K_2AtU2F=3AgJ9JG|gp!zD-G>!we z`ev~H4o5O7$y<4Yythjgn_N2vY#->}S_i@c+h=CLUWI zGusWy(y;-BgdPLC-_cW*pc@_KTHsJUp7mmz={NynP zEFjKkhO8VvrKZMsb@FL_>qXt3^}BTqOZZ(x{K^DB1KpqfAz(=qD18?tpM`&au9wT} z8J#x3Dk@RFJqsZ6YsxqRZE9RTDE(EKvA&(N*hlE-b>fn}wYOjQsB;NGj~UlR%RX>{ zVo4Q%hZDnu`;`0$+>4PUGqafhLR999eUV>PbjP+l@bBEUpB@XfpIX{-AbF~KmYui4 zfD0sg-G68AOxl#ivMv0VXasR6Dm{XrgL#? zbN1PLt#^e4dZjZ6N9OK2zp*#S4=)cuZO?ElJ_)%${~RHs&$acT6%CyYcK@KJSajy* z^izNMx9>9I@9dw#{~Ix?Abh;$Jp%)Px{rL7BcjbqM0z6dTPY&U-yit)Co~zLyYN2F z@O-Rv*&eT@toN1R zy&>vApPI~Pv{*k2dtY#^s0fNzj<5)Zr7$ zzltWK+*e7->5go^Bc6)61x>|G`)>-FOzI5soWMkUg#m?zI+B5nGQq%a5%GoyBqElQ z$3rc(b~H}!34TzT0~SXRA*t?j!Q5+z8yi(id)r`mt>w~pG^-vizY&^UN9fQ+uy;r+ z&nAPBCFf8576urZZDZkS1)r5L)t}0bP|_7T+Kh><$CHix4t&LKCY+KG4YW`Ol4fz~ zU$o;x_Agd_T+lRRa_4e(}r2T_I@2CsM`$QSYDd;rlFu(?#uC>A^Op`*h>gW`7Nn-(wk{DF1AW(XYekbytIj z^N0@zBshxBao)dp+%}ecKkp>wCp`7Q0Tr~dhRuoo@txv*!uom@+yX7>yBe6S*suJi zy+`WdNy;glg|J!6!P$v1Q&@$e?C1wXr{Bm4;hV#S4!8 zb0U4Zdy!5wXz$m$cm8AccPW=q z8S{_zbha~^fO^)UBN1Y2h(L=!itK)C3DMYI#@`*Em0YSwJ%pQWzf+a|-oxlZO)1%+ zSP|}-pVA_WE*ZGevZ0qX1Y^E^3_sa{8!1u7Y8VVecxy}+Oz zRF+GbCCxHp^WvMvahTB5F9?n@MD&8xlmThUU1ae(Boga&6GWRcTEnNKnjULvahN{l zFNu3X{}BK2ZZsxn1=tGe425JbShv$$Lr)%rL~r^Jo&2<2^)a@>S4I@#gg zr^n^|&<~5hNimf!{JvD)f={_#%ExmoNNMy3s$4KpaKsBkvv}V{T`s2XynDgYSQ?3~ zhG^SepDE4kCy6x+VM=)|$hNtLz(qfFv;0S9-SaHt;6$}uB@Y1pV@`y}Yn#FzcIU$W5)NtmqPK*x?h2bXgG?0uMl zZRhgt9~9!YengDUkz(B9;I(q&xXqfa|}JV*CqWGncWM4eztsgvVtz*{vwBb^Kf zjXYcwgb#!;v%oKg6Z8X(t zYrTF3Dt9sQg1s6JCIz_c%03Aj^qA4A_MfLn+zu3edsvKr)<8_~>OOP!r5u{4f%+FE z5sZuMov{&N)-2)YT3H2J2?BDkxN3TwU)IFc=XQIbh>mWM5;N7jer$~7wr9cwLnfww&&#x-#Pbi$6^Wv`= z;w$}S`^h0@c6uw`fl_>SFv{)g=S?9A5Ph>S*l>o?_MH~i-lB*L&av%$>M8$rgWdjB z)}$%sD_)ZQ6M7ixA>(~@YxJcav(MUoI3DPXubwMmq9fJyfhFo6NJ=l(F-5?!%zv-L}U#DCc0 z5E~IkTfR4Kf2;ZW+D0k+WL_*|=LzX|MctA(GqAJ^ZqPr7-qqHbz?Q~;MnHGM&zX1` zjCs>Sf)JOs*hawKT8zlPAWqfIfQ|ne`g*fgqt1LU)>Eq4!CRxhrt7!`HoO5uMnoRyqKnlcl&aX+@vHR zF(Rgbxxp~^IZjBiEX4D}WO39UGIqm0I!js%RZZP5`vn2e%=wj*&|d-A}v0I#Y2a##Z7bCnd(~sn0yoQ=_qFDAU<$+WlB}-iTw+Ey=n$kG<|e?$sKTG6Mm|`BlI{6W-u*DC}z@ zaPzsd(T=?pdtU`>qBWsA%IquwI)J0$4S@Dq+|^C(Zc;u&%0`?SC}ryX&fU^fkHpZs zB1H5mBe4am3%Mp^kKPjr4=0Z5C_-&#t8cGP#uQX#~NOPXXvM{6%W~Qf`EzVXjG(-!ll+e zVGvu9V1Rb}8IYtz9cE|-0uM}l#d_Y|vY~!WYL{i5V-|p(>8*Kr9D01>YC1h6Tu!P9 zQ6E06z%?sszm^45#sUGQqtG1vb(?eu`2ezF%NjIN%=Pbzo_?_F@FZtknz^Mh_~SO; zi|i!5+u_&;M`|5Ev3ayi@nZ~Vme$Ysbqmb=PY)C2iazL9C%E=1Ff5PT3&0cy6IxtG zFE%#Q=dH%0cV$Gm`#)$Y5{Lnig@UYkUV99-A{hTawt~?p_kimo!H#GEr-xEi7003~ z3kw+|Kh3-cLdnhXy4^}XsnPg8b?>Q&RcI07T!-A|K8i-a|I3WsKNx&5UG0 zuNC-CfZ5{(d@YxxyZba1>ZdXFI;f+yM6OT39Iw=4yfmF1^$%Pls?}h9zDq+)l?;Bh-NuvkaA6e3bWJyulIQa2XB3)nbu9ka| zypN#mQ9VqT@g;qVFlYpY2Fh^Q=NtmIcQnKC596bxjMBGf!!D#0Z@&;3c@6?IvSg3? z*hk`FWXjJoTX2WIaK=x{g6JCV{T}_7dVF4_v;dy?z{@AJHJZnxox%e(hyDspM)%ku zm^>^GEuD_(!lKEh2c>IJHnt!X5bGu13k`K__VR|Y(M@Ik#Icj+8rykYJHJ?Yn8aha z?yI+Rbb?J@tPnVbD=A4Xnk()22N>eXZH9%-NVF@l6*UZsy;5>aoF`og)6oseI6M(t z@q*I&QV60BJ;rbOAlzPMv0j=rZ8rpFVfg05v7b_*_e_p5R-}EeG(}WYS$L5ybGOMp zQbDkn3u>Qa3g)+c=kR;wuRJ<#J)f=xbutsL%~s>1;Z!n>2v-{vIXD(a_b4m!sg=23 z3LngL5k>vL2!8Kp=tlu0=^>{?1}S8S;{nG^wI+<{J@${M@s)2Mt94T%{5&p8z_=0{ znhbU&=PuhQbPXntQjBI8{4+xbK&W(Zs_9>8&iv6!iP{%7d7W zw*U0J<`g0_@3GG%rF$9Cc23VfX|jlpY1zm6m}!e!K{%L05xb{Q4|>*$4>Lfn2>$g- zRO2CwFjV>&V&u`^d|eYu!;WZnaBc6?m8sV;>4^ugxJTBmSff6pB3+;&a$m=%$t^W3 zh8H~6%xkyP_jJtcubAR87PJR{zNxPpL_M)cW7pQ4(afR)N3kdm1z#4Q819|V_5K+l z(%uMUay`dsu7qL!^?_&tPZIha$Ni2kSGj90BR;__>sP)hiVYfKsMHJ>N(i*M@qFC5 z?|ztyPWuUD#m$GV?=z$28P*<2Hm3zECMI`Jn>ydC4rAPBWUSU}M&51c^ZOSw{mQ97 zx+=>)3##=f*PCeArh;_Icfm4JOn1K+G2flxrq@L^t(JO+?IFD4?2tg~xm)-7I`G;E z;2Cmt;>RwF0ovwmFv0Jxbr-{5lf!D^#F8G9C<%`DW2T5z2_lH(eW>$n#%*!l`yD69 zbBf%)E7xr~${OUH=Uek=W)Ce8#%z}->e1wMdq7H?gwX={n1jiXOX5nmWked65_Q0n zV?oO!?dXbn^4Lqcs(G6HvWg21q*{PQl6&dd9a#J%0@xvj{B+q&qTu^BP)y z<@Sznz#YIhj=?HO_f(X#a!QasE#k9-6o>y!^kS;awju&bDwNaJKT6@ky44YB=U-v_ z=(o7KN+MtUUce6Tiq@xpWBOX!#tHipRQ-FMLvpKay`HZ1==@te=nzM>1k)S8nAh%0 z{AKRUndp{-1VZWtH5qyIqVeDt@5o`=V;6ot#sUt?c|dMi3dCrV>h+)qc{UZ;& z={NCr=g&#w^@+T7c<}Mc!`p!kq-S5|NLVHlXy&C<^$LD`;>R;RR}F9Xgr>%;Z6OlX z=m`W%d!dgXKN$ETd|j*S0=fOOsC(A4vXf2O@y1*}_JjmQ+)${dx&jGmA0nD<0bxD zHF7rS5dGms3r{eA7xmi;={rG`(7ikkCs+w!UVy{&aVLLhl_#QpTfqXF0V5{xh|e1# z{A6Oltx{HVL*x+6PX>bHy&@>kx#-;$cb;NF7k9n}Yoecxnz$!)i{%#I z;(f2e72{UuubAFFOwyE0zZ!?rDYGE1S1Al_?vErNLBY>kh+`RjB=%K_ROHJrBJ6}Ym-#my4$t#0y+KwD z_3`%?zA<%Vb$J|WxpnXg=KNK}t!9e)yx~^7cDi)xX6Rb$Opr?nq$zc~VU-7j;p*qD zM;j+51s8Lz&O}-4$EQ?G80q)&2slfqZvCKDIvLtFTpo6DrZF>Iz!`)hbRqhA@@WeP+(IHLiNvw;| zO_3sI*iOkMWW|OZlDB9?vm`$aqWA$g*yQLMDfPEMML!Cez3M`E!QC<_Y^aX`c8?aB zX8rpy8~UqmX)xnR;z!XBQz@z<5UP{K0{^AUWV{d3H=nPdsJ4`N0Fx~ z={oACHnTGCbp>YR`Z5}mIG@-NzdiMYcV?8)=GmDM2-mDD@l0hC!X3X&t8FXSY7#Fo z8?vMOvTNDgsRhdv7^=_WwcSK*3>-wJqw3Fh{U#Nmwx| zWTI}$KBfv!_puEFY@oO@gcyr_&z|zFf0@_f3DGtu>&qv!8*~FOGP6n!%P(f$b4@{T z9Is!LgQWI&F<_W-1echa{b>7z7`yZB zGK-6U}KZX?2B{CP^j~%;}0F|*hOYLFxSV-Wr<(cdfMwf z2!UvivBi`|3n!Kj+jLA$Q{t}sX9%IglG*ao8p=4l-jBK`z8V8B_nlbmtXQt&&Te!{Ni+#R;o`&baU$(0#%L`{t-Mb*X zE{-*j-8P=Oy^+OU22vPNpTb5IZI>|<6IaK`w;H>6Rie3FYy6hmsz^dIkNW+1>+Y?w z*LTlY68yc8zRRROyowqj&FixI4r~dg%%)xYp=#fvp=|d&c<+kZ$QKiDF4?h=H(kTW zaJsxDz0?zYiLfQcqBj3alEE0~(G&2c*98H%d%p0#rZs3A$?p$830_6(R9-%@`Vt5O zzz@lmOnYDYg1lB}0{89a6%ify3UsSd46i6+q1|I3UuAtL-;4xD<5;`+HodZ#ePp7f zNHD!2=pD2lb<*0?2=zeHi|O~F$RR(y`Pus-%#vd+g9&XpVl*@z7ic{Go^~A0@w<$FWK@< zgJ;PHc|U8Sy8N02vk`nhm*WydcG9Sr+kXhuu2&&D$d3^}46wKc+@L(_VorWPvD ziJyOl{;lb6Z&RHOau+)a`0Of3w=gU^^u*UW%7c!hDs3@BBqX5yvS~KnU!2b3*f=&T z@KvX=7;Ni`LjeXyn@^LI!Ii;efBLr8v;Fj1PDQemxYfr(e$FCq=sRu`l{eV|71j#Y z6{9<31~TQlg#8idP(V*Hz->Es{|J`dLn)+3UXo!@1<4lwCL%;FPEQ8m8#Z9p_S)V2 z^J(wE2qjs&OlUjKFKG!fw8!Zg$IgN2FU(Jya=E->=Vd7MN=wh;^EIl)6a*+xPut({tS1PJ$x96DXim-W3JiEv%|eX@A4Y%N;fY}2g`Kt|_VU6! zcrWq@55}#F3U!I`qK4JqhGWD`ecH4i46)a`NT4Lj(9NnJ{pIdpRs%DSX7jYI3^zF` zen=Elzvxr`G%#OB?OhYnBI%uAV;E+5;U}*cJlu2h1Y!E*Y>wxw5T0S9IML8wFbV9$)kk9`xJUY{??%C+Oc}aFNc!3Dxga1j$h_`UL+Sm-cbkaG5egW0#bD5t%y$j zDmiS4a(nqg6GZ3_NKDE`SjH#On{RCDt_I78tYrqmqXX?AMw9OK364*A1*v_91GxXN z*LB-0%hHs|RTI7_uQFz?yn82gV5JjUE)?H>MvbXhNj|@_pZNo5@zD2n2>f@Mg1APf zd*R`HFz6AcO%l!nf<4@OnGr!agsYt|9Vs=R5CE1|OQLuDxWJ<>f1eOI=wn$A{4waw ziFs>iZMgkUJ~G!hImm26-V?Cn+DBqpa=Id|HDUgyOGOQYTK8N6klmQXfuPdMC-~h| zvYkoQ)$YTmO37baL7efb37lEsMjy5NFX)k_>R`Zq-u{NXQ1@(y_JAS9x{R8-#S2r@ zc5|}4=|rMzz{Sq{hwnwiTZW_YB#8Ws);+QlEMq)v@_UlJwePw0s!j$*l@&@dCGq16 zODzI@_Nege6u-~pTa@)cEaBm}E8&kK^9o{IbbGJBi%7ZcVN(B&TJ4RU3tatzL3tTI zH{r1#+rZEY(ertZXs6E1)a`@E!w(7ZMu$ z80e2I+{h`uK3>!dLej-&s!ZDk@f2D9^h}ciLqBol^19UoduUIare)i8#nG=Hsb1C# zFFXOhE9-_ZqP{1Ksk#(!*sB80rD6lF#J_u0s_crBqq(Ir{WWyBJRVy ze5E-wahtq9HJzl9M$dR}_>Ss8vBqyO?RW5enfcFf8l4i!0F|rz)NL%@2}kPn=gh_u z(0=TePfLL|&ppAJlfTg|89%0~d)O=@X>@BmviF8TUNlaQwe=<#bVuMtzuh$qWx$bK z5bh^E>Mol66T_!ZU8hc#>e2j{y)+^4&NSzqNL6X@k)4}4$wq4%S<%kHQr?t|78VtU zg|f_K5wirSCeIOd5+;J+ps#5<0UHbdl*DN3F^J9WLl|eOaSs9alS~De_~MOEpU7-$ zq>>xjj2$&e@v zN29z%^riFPPdTmFOLro5>BDW1t3NW2eYutK)Su{RS-wPh4sB!`)}ClW%wQ8d(X zqS`_j=3_>D?SvY*t_qb)591YJX{k{@GKAwE((9Qy(|D-rK^xFLmW44vLj9{zJ@Q%Q zLE?$n-vu@xF%xAzl_ba#D}sY~!@>r!C5Zw+@BU(3j9$ zCE46N6&M9pFme%ih>vB?Y6057C#7P=VeNa^(#~b1798mWV{D!l@}YXY3zhSbjD-Ce zXiyDH7?y3FoV*IYXowb9+T7cnJbACk3snxp2fS)KoX!g#<$XGPkj^sH`=+-xM*m+NeZ9pyfq zX(+~_ZlE!_K(}Ch>wLdY@YnxBP5loy>Hn201#rfHEm#U0C0_i$V5xtT>CXRxrT&iu zzkk6}084!ppLgp!;`7#Q)FZdNx$i=|!h*+TcVi(C8_qGM>6dp7V!gH`;tpgck(XSL zczhgn?=JDwE+%?a70)x>gwW?2I6|C1le&cObQ0Nb8|wpySYVZf8Ara{Od+zzz<3Zu zaekr|7XE$+jLG-y@bT%A;s$lO1E*bWtoMOAD9}6_3?y;#IA=ICb^eRS8L)uJq#{b# z7giba5%*8URqxfa$OLGkM&8K(sFkHS$hSCk*PyQEmFl2*xg&EiIIIEOzxDbHCV{U$ z3N=IB)1TobZ>&DSyf(X7cJ@;1Z~xq20mjrLIcJbX>{7!%*2Wb|oJZ@{O|*m?2Mw=Z zBr7x{2IP`&hUsDDyp&TA73HNrSF17N1ja))j#jrbxHvtn7h{TE$?Gb^DB#|SZ0YoM zLX?IBn96jTD>S4u-F}>Ls)?!14z#E7Q$4W^5dx7=`|lmqJ? zx0eOPP^D7BZgR<yULPH5^N8G@>6uwyH*!{yg`2Eqi+v2p3tc))~1l5Zji(N2xWL%(Cd4XPA5LNp^1c<8_V(Wb}eTEvIlLBxVS>%4;xs|(F~btVL# zY!8X(gg&oZcVe!?M>+eZ;1dwv-tX_X3gB%FkXGf~ZC2Etx<{icB~)0keCb1jUyny! z^=s)?4nbb2NO93Mbrfqgd%G)SBocg+J*>#aJ&5YUH|kQLHc+!XQ(k$vx5jhaJ?5Y#~|%* ziD3>pzyj7~AsSehY!&QveC!4ud@!#k5V+*hF4o84vR}*c&@kozZq@ragV0gI_luE@ z#v2zjYUFMTrqi}%P$lDC2W|%;DLd~5tzHCf3@4pCTSr+aOHF64r8b|<{VPAB&kCXq zTq|qcB5`mPpZRQxFc*?#PJH&o3$kYIRwd2MDi6I?#$3sze2Q?fByHNDbuJwhC?OJva9rS%? zl>`pzu*uNY@o)Fj#8l*{Q~6Dj95SjaL~=FNU#Gr#a=k7$)JBGg+gOmn(jlEgjKR6r zJtG$^90?&hf5`jMkZ@j~uC>uIauLtBL1_QVc?#ptvwqZRmv;J^@$f zA55bjAxF|$mi~70;yB~wZ9G_H9U~xYy0@{qoTUe5HbTo<&*yh#=(7|T<3ctrtwMKu zIhJLbx5+q4?c-ad_4W4B*@Gq_9e^QusM-(}{fA>1{m~>2o8MLIkYF@|MyH7Z@Av!s zVgMBDlK`GUJ($_l1VKNUD85Vftc-@VuWvMgSl4f<7PW3nYy z+l_lB8ASXNeBOZv(*4$Y55I4$IhaB*E(Z*_FVs5Cn3dPnQLI7%!rOfC& zXFPzAXMRJbQLo&k$fxUjQ6D>Ur7(r!wLX51s(+#K*tlG{(A?69JOxuw5tO1x6c;E* za$rAM?TNR@aXbVD6^DvT%t(0O<}3DX?{MmW#s2I6+Er=8djskF{O8tGPS&UZvmg^? z_Vf8WafdjMmBw6hWN9D_pu38?>h`bK@!JGR9`T!0?*_U$y< zu@7hVie?Nmil2xmz6IiL-CM|hyJ~C;Q$8E~6E75o00i&U3KjU|uq~=>V6+n*-B^5B z8-xlrRaeeMX-Z2o5x)zRJqK4(;5Rxs#2ehm=7J?q zDAf|L0DeC}HPt&iakD(%hX~~d>f`#CJ36uSds6w@)w72 z;~x#Qi|q>*J)Q;c<`ISA>L)CM(No+)X7tgI-3d|aMVkhde<6xzI{OFsy{+tein&3H zpSQ0YlctaXYLc%`^sRn)d8vp626WB=4zS8@my)zSLfKu=HGwpok%|D{m!NsX_adw+ zWkV^(oKY8*H7STAQY_x}QgXa4GUav2nz;1n8!`SmVHiI58CZ(G>SutyDmX8uqLx0c zcM0yxp$Ah8_2D`8q%*>)8ce|@ytsrJ;S^FS1va-sB629B4dh z8Be4^e6VzQquPIv7{0QRWRn0|el-#vZ6J9#A+-yH9;(WtxtW?*OLFW&nKJ5eN3T8^-D+^Dd75we8$u|_y+Naow0C(#kGM? zO(xqZnjeznkopTJ;tYN6sLVP>be7q1iTzXhWz+GU-xZzz+O4L-_D=8}D^`_0L2J79dZx5p;AgdiLeQ#C!)?Qw6% zd(WKLlY4#e$e+K$60B z04;DMvuBTe9}>K*?sl3!%qOWzLXC@k%R$7+- zf#d#F?o2ZM{|_zX?=rhII$Trrs%$7Ed7o5wx>=JgC}%)G6uK0ECc>C-9T2wtA>5cK+`||paIK! zk4LisK?&ka$NWNK|D(7L-vf*!Wi%mR?>1aaY(H07L?P2ZzlniBl4C+Quguw6AFCy3 z$*=v>`FYnxI25%8%;V*WM>-!2i^iJNOhb;N7e;D4=AyiCIPS+ip(fjGs}vQi7q^*D zN0{2U!$8A&y7w+QEmzQ0*Lrnjb3JX(3%Uz@tHIVoeN$|ITDT@B^?038 z)}!FVAz~QU3xA?MZVWqwvg-~w4HW7p|ps!tu5uhq=fD|JD-h~29cPn>G3prxg z%ea$)o!39{+5mJm{umzf{2Q&Fm#hy@O%LNh^c0{vRanOw9KXrfLlquxm|t-O{q{IFcxWk zMbC(w@EcwPX2|cvMxOS~8KoMap2P=Nyum0DO7lD0HzcGkHJWx$Dga7K z^gvOv)YB2YIhb$JZNPx>7CxD)|KN`6d&r??M`+?DM`53!kiANNU%SWa<7iwp>j0h-1D_W_d(zEDr!9 zMZR0SD&hWxQ=!8#jH_VLGt`(7kYn&)8ds?U- zi*3YPjxL#ZGJji3HHwN(Rsf3&+wTIU_ngnIwhBa?Wx_xER@nWZPd5;w_o?<}Qrs?( zVJ0oRXo}R`%Fj#1J-X@d;c*j1 z=v&l*QCy!n#1cR^XJlud?%4!Ua}dwrz%VfccKA*RBnhm!r=S+=XirG8p4_K6N55pI zL)~5Z56yxX_SFo~$6SwIIwF{@k&Ryz67lw{)=RZ)eb*WS5vgOdl5fX;P?7j}*!nu; z_1C<7CU1?pgdrbj#z5|{EQbSU-P4@%1VIzag7^4?{>{EBcdCBcUpnNd$1*#3fu0>$ zlCJ+cc#vkb>5a?f7fbTT%e2>r-xjl&o67gyTR+u2^w0fSnyV-p&Zh=Mp95ur^m zDf^CuuV~Iy^FeL>65g>Tgev`BF2baP>vE*+Bgx2aS&U?9(4zEr#3!==u^1Oqo>vLS z^7MB=&E+F{OK$ZXV{aK;bP3lQ{-3b~RE_Z&UsDV_)Q znq1BXwAD5~R6n_tI1fIn-w+t}wR@vMkaS3hdc!H;-?#J$in)SjLRGM^x?QapjdY*= zF{cSwti8baIbZDc{3$c7<2{H)Re6@Y;jq`=@nk)(_I&r*%H6kjol&z(pAk+)o7XWi z_pR0U?(a|W36|P)5BvMWvWME0Zo|ab=lZeVgL{PIp{dT~g)2flm9X(P20%N`lpnJk z;p#DQmFM3X@OJ-%TL)c5vKN#y z%mYkEl)uXV!1{}v0Is7^7CM54P}J(a=_QE9XUfNu1ITkl@9F72l7W-oX~OqB*5kR6klrwr9TWo%x?R-0u$Nq5s@u*r4$|Ur4<#lhz|qPA zaQl(K+rj^yLIQ^DaD2L`yeJS!{8QTW#a4x1K0nVkGCB|qiC1?SUQ_{E{Bl+djLSx; zgYA`FCvfy->HxttEb0AUO*VX9W28+yBgcQ6h?lxT{Fs>r-OkcxLvZc`N>A}2M{j&^ z{0T9I?U;N^v+83stPBxaT>rLzMKN_G{ptHZw_Tl@kMC0*i)J*vn?$fzqQ8(7WJ!2= zX#=!hfzVHhL=)wUKYl6p?XViZ1gNoKi@#-1AsRFjj!|tY=)i`bY6jJKC?)#!+t0KI zC?P+hg+$WD#-F~^*li@MH2WC5~q+R^;v!Sq1kOa(}nl0^%ew-vy@k2{fv}r^EmI?nt zNQ}YVhJ4*`U(U^b@;+|DUuX4n3+~aNn1c_EVm-nGX^J+#`4YtU<;IPRo53Y|NgD7! zh+gM5=6#BestsW;(@sY-AUJ_-e;mb-eg?q9VBaas4%!+7A=+!Z>n%jJ>3QAV9olH6t%d%0ug|Cs z?^yJ(XdlO?(4TEv2Q}9Qd!uK^``t9qb!g4SQK7`Z$ z>AR(B8gd!GdFqHYf@FvNVEa}3D`(EYnD@4R{9aCWjrO5%^9%wp>?-0gC8{t-vg?Ms zCU0-8^V3)Hy`F?p#KOlTLYnVfSdW*bvJnmnk9s=*HRMsqdU0U)e$cJB8V{sieg~rF zY;@Q+6|@zfR~Xq4VQdIr`}qKm<&rHhhs94Qcoq}7^pdt9P|kVq8D-G5p5jtE-GezL zhn9GpEM$8`>|;#V^R-AY7VT;Bl^zmU z=H6{Tqa4un7bn5Rx5<-3?M$sNa#3PEq8L%A&hW$R86ryUdmdshkgZVf`%isOh@rR@ z%FL6CgBBl@j~Uv0387w}?q8D2gOzDjPB{`su8w~jr|_>{lBo7nKB&g1e@F7$i7T+4 zb3kf&hW(ls+=c>vXrBu#PGt?Xf(#kw-Tl_Nh@6R+aoXXV4q`@7p(6Q}AzKDy>&;u$ zC$a(()%qf9KvBB6$nT(JeeLOPx)fBtIjpVC{*mjy*1t?l)?c&P2zv;6QNMr zvn-IgsqPOcrahl$V!YB{@yJ|Oz;v?xQ_sAOsG(LhzY{aHNXAO)YZ*LE1!2jB7MCN* z?BC|%Z%E+149gVG!pku$88i329JiZvK^?Ti?gZIc*zW|lZt-O)gr^919qqUiQw;lK zK*(*GigpFLBOGg7nFxZxALIXl>v=eBHeV}eGVw>bg}=>bw4B4(cae2nRmFKfOR=aF zbFy!yV3^plG0J7>4*6>H@_M@D9MELYajpYx{b`c#k??4#BybkDsOx*C=#S3WY`HEs53m#s#q`Sy|ktPl#9Z=9iP>0hHI@Ph5K|f1&BH zEhNs^8n#~WN?}4rw#dO@|i8jlHvJ-?DgN&MBPQwyp2@7+Ov%Hi_5y$6K3%#pHSBg~k_neN?kTV?jxP@bJ7Iwld6cxw)t-pUMAI-rfL(_vHT z5&9XD8$leZ8*q$FjI+miUwayBBcT?z(#O70-XneXMW9hD3eIbVzL(#(E*h}oxxnas z#M|(f8k+4Nt@@?7@B+N1hxS;2)N}EA6J;|&B7I?Nw`d4H2ewPZj@W{|ZxSTsAsSWc zp;Po(wNEy}16Hd41o7}vlhBm@3Q62FtetM&o;7m0#d7AzfA^tKylh|i$xwe*yntNG1RTP5#N-C3@dsQrHqQfEw^%5M%=SCD(!TBeo=A7K zS4R5Yr<_Bt1oveczViKREJ`=62^^Xqgu?ZM^W1>rCxLXj> zF-Oq}opEP9d2Wa`J?*$f>5P``alXSVL?|7Re`$s{Wg2%9UrWvhwVPQ2L)Sp|eo|R& zdCEofR@B$_$E)G@9zDr2tQ$9vZ}Lh(kZoG$glV2v9FLOkoR6<003n>^!mgp7-$J`V zv0f5B@qUyEgPKKX9p$^o8*>E7P0RIQjIQIooQWZj)_1WVq4b7#ZDl6wxly?FR#)?#zz^Ss}M0gjgNBW@tL zh3Q%jG#DF+i)7s^m<2VIJUyb}s6c@L(?$w+{PaI( zUY|J-97!Qh0}(u?-6X;rcH^VhLDI~qb^vYA4XNmO?gU8)%R|OHnBew;NBxT7w0yX` zr?*T66=faknAw-yF|24r#Z>a~%pnmX-0tf1*TQhqhuVnV-U+qOmTOr~s4<@h!gi!B zrIzp48Yr7O;I~5R-B?C^3%ma7QqL3;!A5{Iqu!o-{e=dt5}+@Pw+bjfkF&1W&Y)q3 z9t5KEkFl4*R(O*SVFo1KK(X;EEZmR#wQdp>rrndjO4(OrC&@5Fh(0IEFENIEUg(>u z>K%O1SeN+GK7SIm$FtYf^9I8qd(c%M&M-j77-5>f&38 zq<$#elXb-D7J>Ec@jC>kqD&tIOzOQ_wTee_S`|cABn(8%mxqlyie0tEaQ!SuIK;1ki`Q=>6u$=f`>7b>$mDQm zo*?QvtTaB_D;73QR%3>pTyDG9)( znfzrH2p2{zU#Xf5%YGBP$)~Mws4+o>#19GDf^1SPJxh5C2%LG~Qka{mGWattUM39O zKh>~ak=GT;7C$W&2tC5cZPn*GH%88o4MIyZ8v3i}CbGEFy9JEkufA~K;GHwxzWbat zW0K5`)-hF`8soHNq-`{@rx}X3dRJLL*CQemqtnj_9%#j$F9T|lN4svv--+!9S8a4v zxO>5KWYWtNZ?KWQuIqk@jOw?>;pb{1$R5MuR~zJWUfCGe277Gz4gU3M%=LXoCui>{ilRh{w3puX7M>VmiZS}=z4fQj z*E?EZ+{4E}AAQ25U%bvdoz(xMdWebDQg6CPbxGYvezP1A%#t6E^M6}ExYN(IoHZx% zJo}F6zJ>nP4ai83vGZkh3_5dbUY0NP4UGo>s_mH_@m#McEG`gleEs(Dy`Yu3ZN9+| z))w%}0kyWv39T5#$b9Gws2aM3QqnQGLjl0=p=t={dV~W+-LZSB1HK2fw+E7Eaj_7o zTppcry`7O@jLd!g=V047I=8a8K)xJ2h5#ff)nEMZXPz z($r{vNN*3>y-_mW%qjW+}c*FI}pz zX{7l6C{`9rAo7$Ts8+hNCNyERliN)9FSIAx=qwVx9g9y=f>+$T0YXiWJf3go-&Ig- z9MJE)p9r^REfDQ4*X;H3Cy?gzMn9W-&6 z!7hs&S2#x}uK`&CuDRnt{{-lRGo^5X=+RW2^EJI@$WR~Ay(DL`+`VBGdpXNZu~}ao zrtwcL(K^{FF-}fBqrx}o_ea%_#ii65jKFB3Q{+l19@hkEZX1;d#Ga^S3s3y#XOGRq za{n+inI%-9l@T^9g)iAQfFclT;_=-0oo`BQ0t}7za-I*Qxf1w)tRhFiPjj1ym%bSS zNPeT)p%?)9oBau%r6YT?EK8$boTXfU>b4bTuWzHggV+OVqFOfMKy4HaJflD5%Xvli zS-;-EETPFG=)j+e&d^*N*l<%x$&~HX>qA z-{1`sF_uL`982t(C*QTd{X#S_-Wc7KIPH-oe#QT>bE**T5TE~7o2{t*Gxm@9R{8w= zkDYBvku<)^{@Sys9?$uIc`=#(c>(ChS22&LhEQ$Z!Z1vyGR>FTb zTkY$lBd}zGi6=&PgjI%HpY+?8-A^pGmHL9;1o$L$BmOe(Yny|B(gd-$hoc$*lOUIK zIi5P<=i>3k8aiy4NqyvYLlfl#-Taw6ZF^I-oU->A%RameI0t_J3g7d6%B?He?12NB z=RbDR9+%bp@nWC0^AGNA@PkEp3r<44Etb{EAf4kiGU;0pb!bJ}cLrcdegF0%vXRW% z;f1l8_gG_8Ml6<4h-+H}mPnJnVKcg&c3GJi7PD5dbWrKF>)E<`PzzmlFhssiAmr3g z`w5Z(o$|h^vwyRFtZ~}tHtJ1wIY`N3eYAZE)+V;>@$!b4W>0H}lh`FC@v$D1 zDIZ+8G8R;FG^TOU?tGvI7q-?H2Ssvh&5^L+a)E`rP7~kf6idPHXZv1x9b#)5%fR#6 zdP9?IWTJ976PP;>srT(I<$bT6?X3ErYt5qF<*y`-9SS<|f%~1o3G#j?2q?Jjg00;d zeAvH$E&gTSgEx0Fa*vwNhUy!>AT+lK_!@}hvF&N8#yEO;*Rz}8jPBo*`JZ-+Bm37* z@yB_8!~gkDM~-@93vqJg(!h864&F-Q9azEQjDL>%+<)*L{Qh76BqV3YIYy0m*aN|X zaO&+pj{Z&9zsBD^^pCft`Oi*-b92bcqp)k&6fD&pe9G#V0jX$Yb<_SVcmL;4!l7Y9 zV&D2J{HNpiKmPQ8IhMaqI$*;3_mMbz9!|L<`@A2+fBuPoOz-zHfKSl>y)&`b%mo`N zn18<=$hwEOT8NQ@>D9Uyvrg6f@@8?k=%0${N4Lu3R42(7S!2BUBQrKj9`Q(tX7qrt zg%-o-J&O@gfJ?eg`Mv!_d=*Eqd9%J@bLu|(vS#}6nxl4L<}Za8cFz$yoRi8;Kj;>4 z-9QoBuL*PzHB=3ipo0ak(N8W$oe;0D`EhJNywO`#z%agF8g%@3E4fo3^Pa+``@4^V zWiY5;$=jph^>0+}kN6PEAq)rW>+8G4lEl9`D2`q>9;<11Z-FFV7uLgr4pi^aIiRC- zD3tMav<;Z|2_ws=SRa?TDxz;h8-^8e9hFC>@%u_8Xl(@?lwKb!#iC*kjD8l2Uq7iA zu;I^Scol98%GBax6b{IpROMaoW5lrhw^vZTG^@F`Nn;Io@R<5|AG8GRP&VfS1^+FU z9PdAVAtSC%wn#A3t$pmLJ%Zo}%j=*|p~r{uM5HW`J>Y)Uys&Uroia{~+owI3bI{@# zqNv0iBC>E+p41?S%>(E+?L!O<8c8ei^{cey;zOVd_APOFlvkVY$LFy~&Kya}Lbr8L zTp0VK&)L`^thWg#R!ZKuN>E-}(m&5q>E9BVF6HmJ%~2LdggGdt(HiM zln;!|1i&-*vK0!T%i|Lvr*~;$oP@tT>+e&8(z6c(`zu!VE1^>$&RKiDK1uBhL3`*O z%NF@m8V!{md+{IWhV&X)o3fZe&YBE9C;oKo>g8yYCmiZRU&Ll+Z&@=4&b$(~4Lb;0 zRepx3^7>|O0JT9zrYsSKGKBuhjXz$t4TGp!3HKUe%sj!QBRYqVnY4oIJu5125C>KP?y zpkd8!ApH-V&ywe;4n|M@vTgHzo-mD8bJ@=*J?5_FrKt69E7B_qWkx?VH2m%p^;8bz zvOc*ZEG4N(h!6Y7{uogXo4HHtW+Li4>Kv{@n!SGVX%eBmXm7Sh7{ghW*BkFre6$d9 zfwEIefpmn~eK+({@>utE@@4FE2PRdOnx-j#Mku;$kek7|gwAVxl%G7w`J#O_@WoY0 z9~TeJ5W{{YK7ecXH4lNS@k2}QK0rfIetPoy9&i+U6QUw1bc#WFBw_2a{=RAm2JLp$ z@pc779pe0}JfDfWFgX}eEHb*jNnxiw!d2|^RT|nS8uJe4Ni?^Q_VjU;YSg+DzGrXL z@@vEICKkSZdtO@z8TRXp2ND3*JLg`W&$VvI?Nfb{SDn;sC%Y57kLrt0Ojtl`6dy=u zG{+qNBrk*MR|$KX*C;%A_1QXq0f)bg20EAYF*`U->h!1scx!dz9~M2p?x3uz_YAw_ zqXYq9dslB$|I|X2e{=^E13l*GVfmcqItz%cGS4Vj`F{)Y8~L3E$PR%%U(=@{a1W*B zg-Eo?6=${0l~FZ8enW(Z=O9VQWgNDGk!u7}q)-UdF*EH55W2ZKtAO14pEg3>dM_OB z18JVEz4fZXprIRU%Rym20Yq4w+?)M1_9Vb_e4}-f_McFz3&*d0KyIbId-<<;@QF#6 zPi|NXgM|~O;pz|GiPh{}+??a-gx=M0?cCf^Uxj|Wkgj)H7n(|<{4r+xu=L1$-U>+7 zM|PqjNOONm$?N_60KV=i9@{|afLsLs-2RF~vE0bf_tNCcU-WOFB`3uWKezBCy?8xl zkRFbsKCRFS$aT3Y9fn9#_X3HiT?PvMK|mai58?V(Xa^ttojrW?SquW{9$-RxKl~nN zptH~8i8)k1fG`>K;=MHW??4&C^`oHJ>K6(7<^XCS|pM#0|qi99j zX>}a{2H39{2E{cGJUF-i+40)f{H+1Sj{QPQ#<2OEpj>0Yf$NjQRTD2;(WW$Y1xJ8O zHy+EPVu^d(XJ2c-;7DXg2Pb^ZAbVGUYQ5FnjSUS!xS2rrNzB!?Kd<$3c`i`7fi)~c z9wW=e2@SdE)&8|;AnGd^+6(QM@muM1_6s=D?SYbDk0rdOE&aTm(RYX2@7?UBEXAZ; zyT9Mp6Lg0vb{Ra;A~t;?+k=uDSLhem<*$-utt)QBqnL4_BVeS`!CX_1r=+V7@U2Fg z8<`NKqCin!Kf(t{nbkzBk0i{%ZrykDr9AExBTnOv$?^R_!#SSc9zVe`9D9OIb?^_) zETtiY63Oa;o;8CwJy=&~j-S`Q(F@I(L|sz90{OZ0A(X?MwG7PB6LFM6`Q8`_kCPjU zU-`brHGXcNT`T!ur33@7{Tw7!{}-hzv@=U7Hu;#`y^{~kAh&HUK4HRUCYVwEj(eK( z+LN@U<}k%*+v}AEm%?9J|EG7`!}?>^JswuLQ- zy@Ho6{`iZ5z91*Q3Q(Vw)a|qN0+Dz$Jt@gX7Zk4jXY@L_b(w-6oLq zj&(S7@Us8#@xcmmg8gCbwM0cz{WV)7PEmr7n&PfLAcF_zzN=FK7r9SWJ&d{Wy*s*H zuX#IX+bOsz5L~Fj7nNXOVtrq8DD&XO>xaA&@3G3$L*V|*{rNcAT%DS}cp$^)Mb>6o z`vjrHJP)3_2RkvquM0o{DS&+=x*YDk4UL@So-32pq4M8jina)n+T0sdVu2!i3l&PQ zIf?-UmH4TC)DrjN9-SIK-)8e@>r0&#AHRmEsjtcvWV3mV-y37gG45ZJP)SID!lm{< zm*A_`8$|oqjac2-Y~UELkwk{2!PvIw_#-t68=Y4eg`b)@nn^4ifoTrHRnl3xCDU1T z9!#aU)W(q_>^6QB_wG($tacnxJI3qNE5C9a%nQBWIlelcBlHr0_#ZJG;Okr&1M_Fi z?0$al11J%o7bmc;I|x<2dCmp{y>Y3p{wf#y^HI)lTC{zh`tCJbkw?*DV!frXq8T}V z`wnCImv?`_%=C^ZPg;#V`~=w6{}kcDyic9u(xA?kdnc6VfZ7r|k8Luapr2G98Lc&S7Zzj4Y4?lxjjm93kl`MX2!Bgq0iKo{QV=EffT8C$9>EL`IG{wP@@MU9Gd3WB8_h*B zNe*GG3)}6fzkeM(A@Oa9Kk1l}7H^1u4{#PQrDQ-~hSbLTgkuVgahkR7_G_0vP1kTV zQso8@rkZ(oW={8M+%heCsZTkoeAN;-NgAl!WedL4h9tXWWieO{B_r@Q5;xUtYWK1by)!S3b9zU||gU~+f* z9+DU91lR(7dAOU;z_7>cfXhu&MgET3+ev>uk$hUQm?CmkedsUiJM1^#!O{SjH?@(l zLbO)bmbLDPaPdcdUQ~AQKeC_yFzu_y)erJRLY9a>m2{uY2o*%d6dnR=vyTV3j^bk& zal5jRJ$(VNY}57=YB-(>!V~z=(Z}?SC9W6_RvVS+>*y5wSU{j^>udseg7H2-h`)V+ zB?M@duls{lWU=8`E14-(nU|*$6tb~;T{xZVixeo}6clH^<4m_o z=mM1}W6mk;4mBl==nlM)@N_CtkSDmW;7INy`sAoEQ!haCrjOoSPE)O$?X6aVwYF(Q zRA6!p*8c3pXlZ{H_5puwGyW2w!S~1332^5(xR!pyE0g<~{|dUAUnM43NO?L?@4Z}t zX{LT<6FLSBPVK=mrqyTY%F0WQx3>nPSSb3dVUo#b0bv>80TNv+ZUEKKxKGE=G1$-W z+TH7)0B@gDxIoKY=?l#43-+!Q9*J3Pv{xG5ZU7VoFU6*xh5iA?Ize6dVB9t+McnY< zn##%+BCt&O_V^Wyf9tN3oMA0kgF35HErT?I=>t3u_{tEkcQl<5bnWXfdr{`pJPOfT zI&BT_xS>VkFp%|vWdl;gfI~3Aa%g$R*R$a@=RDzzzC3l+?+lnMtu-!sxr$Ar+S9TB znru)anf9F7(T5f##C`=WCnF%God7QCf$h=4s4v!OHs0|QeobeGnqH^xZ%Owx<8DKx zoa(z>a^{SGA|9WW1=o|=ew`Nyt5ewU^tgFP_l&z3Po+aV$Fm2N3G3ARq@w-aCR3|Xz8P+7i!!IlXe>_4? zc#%-4j;}WA(9hT69OU4;#D7}d???S=gR_jEku`%2#87sH!VThpOYacNUjre+o1;$} z5s%hUh6g3JT(&?fYYw)~5UvMo5i_iMKQo1`Yx{I$`9CjLNsv4i&Y8omPK+C5mj(-{ z{{D5l01jHxy4-#f&97!`Pf1FdfR26W%VBE344|dgzEq4-?`L~mxsOvAS-8BnP%eih z((v9#9gbCDUDE~8MJ_1_Qxnrj+91=2YgWD!PV>;t{E+x%w;#dEJ|?d*#v+*5Swr5w zsea7EyZ)t)>X(~LvtME4o{m_hk2G=iLOF*;Tdg}Lb7QE3rWQ}=CG4M?acvYTEhzgr zcAFu>G>5NXKp_NsrH%fQSFWzR#-%>CU7FcPRPVos48@fu70h@YJ+>WiDd)k{hK@5`*WDK76RzS7jLrx z9DD5UbGXuI0F4t870BH(K5ZIMGkL!zYq{YoX zzAo(HTMM^FdF{dach5*+w%dCHS#!`|%;FbjOMi!wg&Uj^pb-&+rlO%92y`Yhwa4IBYTFg=)M z#l`LS2;>mvm(0bm2U|q^CO6Ffb#lZ#iuD58w1d(^~q{Z5~@h$car9{0yXB5cU3&f|Bba{5h<*$pNMd?WMl0i^E- zNcx87KFC$KaQc}<`i0Zn+(ftC>Irfz^rzkevOB)IhdC-YbN$xmo!2j>G+bgnq*?oN zqSRoyd_~CLn`$8bj_(FAB;fQ6`_|ds zYaxmIMd|r*k!0myk?2^zG5qV9|1K#+Ak4RNt}Bicx!Uf6p;UarmLBFO!`<1u{2;UG zf~>$kn7l@GArLtpGt&Y%;FBy8se*RT*=N&B*XS@kc-U7y@qpbmLE{KmI#B(FCAOWR z#j6cznSY|wgM_hXWO}5bB>7j4*K6uWMhHzRJjnj?ECM&s^2P4?MJP^4M+6?B-}8&U z&t5`uy1LvwpFb>X<;c;Gdj3xNr>JZD{0bou5Jgev{uGM63f3FG-5#+q1(PssX((UF~MSbLS)j3KQ)# zF!00=|12 zc~tVS!wq*-8lAmxhc*;}HqVNu^Y_>oA=azmhrUjG^o$Li)##7pHl`04U0>vV zB_l=yTxfQ)aIxtVv{|nGOx1n0xXbo!U%wju@>qKVE0>?i;T+J|5E|7t9==%Hhe2dR z%V0V;7jDeB$mRYr&G%IPdb`34uEvax44wEso|)9!yY9#5M?bB4Pp97;F2ik6u2%65 z0}AcKqw{sx%>C^HBDj?9{_S*h#*qoFH1Yx#BrB(xpis|-e~@=0^nceLAR{6T-_2)5 z`&@d-O$8>C)Mr+U52}QZtEBf_>~#kYE}i2&B&h@T-fmA8D+UHC?Tx+OBy<2lf6DhR zf01J!=Fs*FF`k6Ue2eAF(VfSA&2;fRQ(y2G=fFEn#9Y1b?1p|b!oEGd9N;!!b#_@J z{QQ7>wE{Da_K_NPdFZEk6?<-`&Cn6Hq%f=|;HcP;C!VS5S`7??71m!(y-{TBTR>>0 zazfLFn5eG#7%Ka7{dS!lCWQwTZeMV|^mq0AEAIg=SR8n9<`3ws9bxj45+NgV#~}w2 zWmx(fm$2daW}xlnaRkIj|CYaDElWWQ-u#7k&|M;B^!MH}y}luozFE!jbHD#2ss`id zhIWIqV;VlYPFXG9k0+DmVz@cKiRy6O5`#D&>aOxwzqVIB$bMUXCAsjh(@Hm;(&AGU9Ro(gC8HY{oo7yP zxE~(#BmdqkfEENGOvZ3u4~#bhcOBP>ksjw5F^Lh~9@)$ZTZa9wJ`KTa8~B{q&RCD)>H z+>=EG&e}4E#w?53b1oFfNQIH2DN!t};aCB@WS2+0`0UhRSW>)onLHab+>PH84hum( z!{$(-mL`?lKD?^jq0)D)oPx%aq2+VQPp@Vusk`U7av4PQS@J@5P={^1q#rCAyaPka z2X@=D>!WW%^Uj8gWX@A_xC?M!K~ePMKkOezmfmw;*?q-wU-&_(AO^O&j|TO&AIydnCaK+J9%g4FVj4P&#pcyIohayr5XBHNbDiAGbTN!daQ(-W z@oYD?yPT`K!(2Az>yjvsp2jm&6f{FpuR$&oexFkLomGm?^yjPU_ZfH`>Oeb|ReVwX z-0u=U5X$=>7>{ZXpK9KyMN3`c8Mf+NQFa;K5V2KA1>yVjz?QI&lwfEH^VkpBq7gFB zR75&-xx{Fh`Ad-Pv3t=RnHGnAr5hFBoak_w%}0IFY_AW&A1wDRDb%mlP4PbH`0;O9 zRKhC6E@zG;c2}8s%$%W?KaemSa`*fqB+u(#`*0fTirK{1yAFt$w@;ouWZBJ&J+Jp? z$|q`@u$>No1%7gSuyp-4i6WLF+QO)$P;J}D%o&YYTk+)pgMbQkxGA7KV(a|(_w;05fF<1TM=|4skrEbwHyXL;^v#}%5)$&^?4gyGb!-#7 zuQ3RoFTTdP{T>dV4RvaOT#?&+AgG3BDn6uF-a1)USkwxH>-<$Bkif&2tV%DotmSQd zhcv!XzJO9)iR<|0jQ5X5U>u#`#=MRE^GU8;G5Z<&M$PP8-{&JxvBe>~Ty1A#qy_56+(G0UH?)qSn=3Fm8_9Q9VvM=&I zyb=JzJF9yZ|NYm!RZc>+1ws}qsX-pTj!;`R`1`fN`9b0FbC!5bB~+@W13B^))H5VROE zVs#vk=ue^2A@x+DS?%`o>alz8-@Nm8XBSaobpwQ9%LE`QUtz6R>=l?iZl3jPR-(r> z{O)5bj!0D8idn-;#*}^js(a+J7MrQv@g9482HcL&=qM0&@W0K0=?=%!-V=XoDUA? zYQGZ9vE#Vk(Qi=7v!SFA6|5SS^@qLmN`?PBKjaNXYa^Hg1Ko}7P4@n8VCCw6EQi2rz(wqTIOw|K z-FCaQ8LmQl&I<^nDu>9vNFO1B!7I4a z5~vB?7|iblpe#NLnzu1u-XI?2Dtl%iWN?tX@GkgF)PC6a_u-bOZhy8I+ThW(Cy-Bc zwNN#x8cYyxX;N!_(;~%wgWsg;%-2G(JsZjrr#xMU;b2qNo6=q>1sog~sTdrYnR(>D zk|IB6Bl@X_nm<970t01yq&sigYFxzOxI}DqDP#%c7FO5h9qcHA&wT-sw5R05{T8l3 zFauz0BM;?2Ykv3BJPf;j7`cdr-}506!`$&l)@|Q4B#=7iy*rw?Kj;{h*tkdc;1{5E z)*HOMY~b^&0|)>S6e-8P4RH!bL$VU7j}`;BWUl5e5_j;tkS>t+ds2lTvyL@`YU-<0<8OIzaahYDz82sDQbT zc7CIM9AIUn`1WglOqpaSOE_J|-^kImc~d#?4US$I06l%|h_~|+AMbqXYAsZ5zW~dM ze7%Hsso?Gv&_RZM0%WL$6l&*>0{l08io}#sl~MgxMK|(3=8!1+rXP=ymvqR3x|aC* zc?G^&u|G3h8KPz$<$PqQ$FV42q|Fk&T216||40}kyvzf2rzj!`?Z+-KrLUq*?@emw zM#0sVIz9B0)6~g6^lC#P{=#d44kq`^AV&(TYw+-nVk7UxZBk6uWO7^nb&LF^M>l2B z3b&$dE)2RbCxPEn;wV83ii@5+oJa%Sf-`P@?ZIUb!$-|G>#+7MvEucK`4*+-V~f?0 zdqi~Ek#BH<-n^T}Y12MX#Cat*_$r)-&vLtAkzX^;@1BTJ`99#|0NY+M=36r) z7Rz?_b+{$N?_~ZlA^N$R!s6Ai)V|+P+Rsrk5fnTo#kfJG2e+0O1E+x2+&M*#j}Rf; zrnX>tM4=j1oU9y445!Xi!x^t?xX0dUr{_lrlAUg36h=~K?`D!;j-x)&4iFPvWmOCc*w{M5^~Fm*N@dxr8t4hXWq zf;WIQ&KIw-phR3MM7H~oZipcv{iBU|1r{K|hb_G)-;&i73 z-j>>_V<%~@t0`bl3-I!+-Cb|Uzkh9bngiYE+XZ1J5|D}R@!`PgotKbIe*VajkLHMi z&LyIvY14+shbr#Jx`e{>q)kJi$@z))TJs_VUW@w%80BSG^0|*eTn*7(qpzu!Ovz z(#~j;6M8=y-qaFxBqpp&Q}QCjI-392J)bMZ<{n0HDNcmwovojUs)MHi?*1yy%)@6( z%Y2>u#@4n+B$1@{SG%GXP)i<*rv;rPJLsCOG))w&1dIzUJ7G1X0m zI9JIPqo-O#2^jeTIX-m-5J+72$dRcYZw{O!?s-Ub_sBq$hdEp@`y2+#@Io!g{R0=eMv-y-ob|Q|ZpMQ#1YkWa3_^-! zA$vF&-F=?blev$+*HdB{K)^!M7p{X^FI@<}3XHoFa^2SsY%Q?HMJ^rUxvU z6uWOfS?it{PldJ(xn`veuNXIyHeJ_mpD)wxhF+wroxYR47i|IjVxPQxkTJd%`YbLp zfhyX5Q2PT#-}c+&ov<`SOvhhOa&fhA6tZ(J<_H&+Q|pwR27Y{1sZukwzu~S9wTqSPPu1FWdzu9km9BG6&k$~1`^3%-;>j)^ZV}W zP4(NPho;^9WV(AbU$tNi;7$@bf=!a#qy_}r%g;}#ft;K$4&wY`i5M-YKzg{1v z<(tRmBSz#W1qNdmhtwaVB27;EMDg1r>jYa15L+nY9l@=VtY1Emr~5CH!2l7zNMtz# zD%kVYrrww&l#8-t159@|;+5Wd`T??*KK^AK3^BO?Rdsm#b^Ca)-8coGx9#z4ew&>P z*2mV7w!xjJS4n#nR_=M(%{lk`(DW78L{iV_$odsmLCt!}+%x?V)W6sFzOuf~SId*q z)iu|A9`}LQ&6H1oN@;BmfY##`CAaqDb4L$2g8?SU2dv26@m-0C4r97f1z!^`?#DB` zr1wif=S+9aEXdPeLow6jk7ef1HM?O@)TKvOP5THh1D@FVoX&~1oMmP~UFo=fltn5q zf9rjnyu=)H>I>#mf<)gs`b^|J5^L<~_-$Bd%MZUXaO+i5# zNWC`Ib&L!3+wcsE;l40-*P{s?DV_7Gi@fpaUW*rPtcPR@ps$T}_N3kh`sz?aK+L_~ z>wI090=SuPYsrW&M`!i{ej8^GPU*dItnSSms~_fhyD0k@L-PBn{wW!$$l)?KHyU%C zC`&-E7yA0LR(P;DT8A75q%2$#rWEkjef-P>XsvK|OBcYy)&&{`#Efjd4kw=peQPwu zg(r*0qY2pOEc<<|-jo~iop`r*R7uI3?azCBC0+*j;%Q-GDBd`XRS@A~J$vxykG1?h zhbQ%H4FnLJ^s0_qD;nW)>m2gW=02z#?IxtvU&CctjOvT)Gl@qrb0e&+K?){@4Xsow zzQptg^qg6Yzx7l^|1#T&bNDvOub&w$-vhw%eJt|ugF#YLGw5x&F~A}M#c?Pj{BGmpmrH$U1@eIf( zxPk91ANr&snk-8ooT3gm4q;bSiL>dZh#f)ttUh`^KGq}DhmC(!lAd`yDq^@%^C}!6 zX@gfDg$g{jdud3ww7>Fpv25)C#kdg*CkK6D126ew=s+BV>eb9>&IuXLS(dcA0rP_( zlhuDwouP%)0R)@Sdld#asDeam2(I3DQ+;igrt^PN^Gix2iKcNOj6()|Gn5S!(mbTr zcRdOFI=2H~2F)iko+OIS$?~jn+BN}>3lUbrrZ1u>z zJT{&Iv*D!yPfcoxZ^U5}l*Gs2h!Ob48=aEJxGBs@_ z3bsKyvI>OMi1G~mx1|#nhj&8ZFQ-C6gCG&anysbCIfcQ?zTVGgb%*_=`NC{&#vV!5AYuEhMptDB- z{Hq(V!Q+JPaY41y`FIATH;tKCY91Oa=vKoz&KQP$;MyOTJ-~Vm*H^H?U}kzSW?k=L zyCBG9dE1)tm)GtLNAz}GQP$hTdOkQafqfr=)fsiB%a>H^Jl>p-a+jv% zcS4m3EA=~R`op!@EP54((~Ukk(v|u4xT1R$43XE|kmzfRh1yMak9iFabt=66C87a@^OT)Jm1AJ6g@#YH#laqrU)KqFIK zqt@8?d@T389(zo-L#JNJ;4&~SBVQ?Wv*}8-Prw(gc!KwbL5#8v?Ju9x@%z3rQSK%~ zc%LAm+gD;rp`P;(9U(38mR4^6pe_i}D(kp^&g$>`{_TKL{_Kk!-2Su_8^%dG6tlYK z*Q4uJ8!Ecq_sOb4Bx@-Cb^9fG_Vbgygp>2iUI60EDaj1$cz9TxPTtP~?3obxVW0&r zkbY-`{(86Yhd{q^A_bn9_0SwV_AAkDnSVRP;bhU}8u|;D_RQX^AqH3GYWjf*C|iG- z{K}Mt16X}CbU`-O-PRr0q~!g*y8DwYD1*P`R#NiI+5^$tfPodyx%ou+XZ5GSq?w$n zk-Fv4kAIl>wu_{<@`8)qd~^ZiElQ^FwKKneXPt>2emfhd zG+i#&Yw^_>|8)3mPdT3SH#87_-nx|Q-k08pD<5+dI8}44%LY^C3f9o)+^0CaeP z)JRwc$yRpOqH{J#-!cFjFRbDn%+&!rz(K zT^S$Q9)p%XeUCwYeBa}^XhUTWDir;AT=Vm&b;-Fu!@RWrCeG(|4Oozj(J;4MT-sj` zM0W0u^m)Ie5BRIShyBz=rn`bt_X^tT|7-jI=k@~&4! z`uD(!2vs;3`-zv(mlWOAEhc)DI|&ihC#WZN7cx!?=dvyGBNDc+wf^fD!tK<$?LPeB z*Z)`8DUPB3VKx2#5IaTU7QxXIpa14gNx}qbHT1II!(DUUrS>EefTNWyAKH$p=Xi0wT#`^5Y#31WhOs8+WIB&#p>^spp!sOir4Y1Uv8{6G4C-#bi0 zos)po27+|K)dk@xiztQcgU;b5X!x;kyMwcZS8uSU{~!O!r~itlgdaa-N%-or@Vx_K zu;aM^?6~RRa@`yeeI({xR%H@M?9Ese=`1#htn0*i{W`u<N=t#)gIJS%DS0>bzT{ z%HC)9O;}z;U!{h-5sm`IhVI^b&sY4wM{P8CSq+?yI>=*&jCRPp%b6k^q^BOJPo^@e z{_M3z5VUL+zlL+&pMzJ=RsAqx4!8sq#A57%%e0yR$!#BIz1FgUFgmd%#J4gRISN}dy|Mhvo)(`oFO(jvt{y9%KgGN+* z!nv10r(MOTA>MEKGjq%ea!Xx-v&}oz#FK0O?jOOAx{GsqMvZkvo1OD_u3k8EJ#;RV z^@CYZW6AJc*U*J`hIOy+w|Nx!eI2X*E2f@((pf|v#&n^z+~a+fpD>=R^w+uRulF`i z!k6(%-oq85(4-sCZ4FQ_F{cT=KymS%Y@&`(1kEO#Q&MrtZ;AsQg%Ba zUf3Sel#QF@rCS^w;fqR47Ob5eMQrvY&8FoXL^IE;pYn8sOuTV+|K)%xpbG_P-{wNKu()zNk(u7>?NqvggwrSMVyl;Rknh68&A`M$X@;ijQh`h9;< zA4bPMX*3vR{|2pE&_SG@k%ot1)Gvl=Ych`u91WE*OVDmky=X%E{`v!uwb33t7azBL z+EIYmF5C-m+Al*fu!3V?^tTj1{`0)N4`m_msW=(;Tw5LIXipO*{`xhakl)v1b2(=Uegd8AX6~o?L22ewQ2lTOa7U-sMK9n0zud9yV_sW zODu5hG&zI?$4#Hiw$|8+zQT?pASmm)F}`$344fq`#DwPt)aNk~Wa|vVKXO~H_kNX# zjn8&Q@^;2qio_sHvg(_JU7*-}q*maK-(QxZwzma1qWzjIb*!AjZk(Ryu`;sbZLKeE zbeBlz0K0mfcfq>#dX<5Q-&+?QV){@ z{cAIi5`L#-%hyh4BwEKmY-$dxbbVJg_H0$0mE zSI9xD)0wOFvQ2WVZcyYr3}7dp1W3JA?KEzQ(l6kX9kbxkIU(H3Nd+DayU+AEgyiPQ}x3+L1g-Bnf7?TjCBQ zI|ASttIN@>aX;=)OpX-0_&jpQU}E>!n9t8BRw--uj^K5wzMh`NC+84`UzZDwJmtIC zXKwjH^3Jr^G1Cn;qW!vOFY2F^eoX zpv=@hfAq*n4sJp59sAOI!$J^yufSyw8@rZpi;JZsP1`%1`oMTJe~qM+$5+#N7L)Dv zi4R1r5*4BR7@FXe`haxkp;)ASo!o-TG$5PNy&<1CJPxqD{(MyRMJGqS$#*pX&+em} zReW7kQ;6XJKc(|?Ds5|h6ZpZ4_L=65h&&MPM_`ih4tLqVsbWn)5uDOx(6O>pvlpbocLX-3UbZ~l+eS)9$S=I3a`B~Xf_>kmNV zbEgG~dLR~lTAJALtS?Zh?Wq#M%&ojLsn7!b&pkD59S6K7W(ykCzyVmVv-Kvudma?G z>+m@r8nCZ)C6VPOr{S(mJgN5Cdbbp8)AYnh5&ON)UU2t6Q~E0c?*lx4oK)S}ta=Hq zan9M3E%ZK;a5qlm0P1Vp8-BCVMhE>=2Jzu}3CNA6JlRviJJ|quG_7f>kFBA{S~y9o z_xRqCZ?}%)Z0+Zc+;gzS09eEC_cFXkBLTX-w7EAdFFE@hLcYfe%l1b3om_v z7v+1q!acO}FO`*6dWH}W&XOe)9-MD*eY(N@smE^q7XI8b`YlD2vsVz*v2fU2DO+G0 zOua`_;SyT(v2X52MHnSY{_+CEY=e7-ehHw>KgwYv{Zwk?y)Ym z3KN*uecRIJS5;T%`w1WU5MFY{>H9aOYjS%;n>E(T>63VIpIClQ0x_*D(hg&%yEmml zU&21kNc)5-9B%l={+W?x`779E?}#XFE>Fv~sAPY*1%V+i+s6d@%ngy+*?V1a3uoXb zq~~b;tsg17S)26r^_@euO%FW#Xk-@-`>3EvqcYWev3!tqi}o= zU?eFB`Yk{<{k(`@`>(g?n4&L&(+1Q=0m9iIMkFlgCVh9{KIgS>AfaZZFnh4%w$=kj zlj3kdV#)Go@|gf3M^Ifjh|K@a-g))Bs%}gA*PMHT!T1uei6$qTJAwnzL>u_^U3`11 zby~aCmn|J>Elp|7>1+&d7^CVbZP)4`yi`3cHxnajA3eQxynZgXu%K&0W{;13osEJS zQ3Y@6a2yC^jm$`_rY#A%4}$%<6+H)2=%taRFQD7!K0+?yZTz=)8)rwvQNR`kCXsb-|D}MUr>9xX_*w z;)w+F$-J}iw7%={a%DmcQIC6#XeJ=pPD8kS>F_nj8kR zVH_9FboyfiJYjmNs;I2Zyni?@Ln*;hJ9ouv3?0u+A3dtIwbU^^uADPGEa8Vb~wZFSH%G0xo3uA-O zOo{ixwJw-Y_@-s2m3_ExJCglQsy+Xf%{{$jh4_rQ7rrjp>xEDf+~)~5wJLuyBu4{U z_XDLfX(|hSJ*G-@Om8XZyw8ht5U6bzvP)n@f&&^nGb~{fGY}qR>zINribF1*sShNn zA5pzjur8l`r;cAQ7kY-i&06io^$lR8&V$76bRU)ugakPMa!MNbv;}0Se!8XPFqI(W zlsK^NA)AJrdnDj*?8^x5fjtt73k0&~_D5zP32x$uw|f~P^f&d>kw&wXQg*DgA$W63 zKzqP7en9SZ&cp51^lU+HPue7>je%5&`~ywmR}nV*d|5w~294hOhYO1HA!DEFuF(8xtz;?o8H!3|k-?}C-s#&oJ z9VSow>2y{1Jr~3;`UozFCWpoO=*D~}$`7g4ZhV%m+;9CR1Gi6=jSIxiTsz-%Kn!A) zvu#nt*jHk!`iUcUEV|rBhmv0quK;xto`Gi5{-x&0lF=sd%rxVA<7AxzhMoqE=Qg1` z6VVq3KoarL&PkYqKb<;?DY*Ld6$JhVIB7`2X zzGnvL>5#5ZocX;-Z~3WBAIQ4%Gk?(RgCYz<1xJJ){AjZSd+`)jPXAoZ`+Olz;eG$E zZ{i6|_Vr<{NUVi{3*U=39z=I~9{0XmcLNdQSwHEt-LG`=A>2$T{GPgHexWbeU>e=` zWlE1H3b!=rZM~JH=!t*r_3YD;u)oC6;Uq_W7%QF|1k{B!b238Epscvn@022oJ8L`V4NY>L6?Ib5I%td8%VO+xczV(y!qdtXpudo__i0%?N;k%9k zWQl;UARfxFT-2}HPau}U5)Uel$wG&LFWCv1kA7U+uB3b@#>Y9hAp{hv*vUc)npW8> zvfu7Emza*?V;@Z{Zw`nDVJn}HI|Hcx6# z+if2u#IJPOo5BLFqO%vJ?WH(l%8@37kVE@__yNSeW2C$l%uw^GDRk-d5Io(Wsqtw{ zte$*&l`{eyJr^{#+Q!@-K}YcAQ}whLGIG%I@OnWm&0l+D#C<@J-%R*Eg3*V+pTGMo zypT&!BGUXwl#N#YXxzj;a4*NQAM#qV*wl?!6dP97BO?g+N-jN35+vc)ar&g|<%HB} zTs$y#<5^NRzp(&^Xinb86?NGnWSK&hmnZnABDXh&qrJ2q{V?hNln#wJLSTFm5fX8T8r9eGAgl#GjAwCL`J?S&xca!YvJyjUgjnu(|NQx6-V zvL^&}^f%_)IEVXl@CxeF9KCOxLUQUoS!*l%4{-MF;dGy{<#Wdv7~EyD(WC9Cp2aHg zvb1V-zw<-=zG#|nj?|=5$b&4iXMnHU-O2<0XV|Dk-R*$xf@T}k$p=8_X)nM^Hkc;M zhUnd8Zi@SGa#x#qcF@jb9;rD1*63KFOt3};Oe3G$=w zHRk(0u7WVT-;anD!hRI|@0&i{35IOyDVmvOjcN#ru~x|eA3~2bjqNL8o4GhLryHUr zDbW8Q`oXX*9IMm&cH5pJ1eFsXo6tBtKXR&ozTyD7E|PwnI?Hch-H#BYY_TcgC$fi# z%S;ho!Xmh2`$742W2GwO!b@xoP!Dc7oc=7dApWCNC^6YQ&W`XB)cxuyZ;`F(zwN3z zHVPok2q`Jt6{fr|cg$v=Lar7V0ywcWq?;dt51c5)+!fLD_alzpDJZ`f%|igo%!wO) zI7EJ@ZMFcGgA$sAdv56YDf_}-N!M@#!ehOF;b&jy0utG@_}e#E6qD+|9{ZhRGY4~E zU?dmP-o0I|V4wPP$D|2Ue%NA@>=Tr_GKXVO_eZ$P(z5nhQo47ldX||hU&{KlEQhz= zTQ!m_4Vuz3Zzo||Ob7Su28!3uf}}^kYzAaWUohGvX{3cQb&0`ck#BBn92RZw?Wad~ zC|-GmcD;SE)l(=XP+fHwUsbD_nsF_GcNWei^c?|nDa=;U%VQiuLG>!?s-hQ&EC$i6 z5ZjislR><#m9+P<{)}<&_gkC<@KE?W@lA2#_wgl%??)r*okkhnJWB1Lg#WBo7(rrSI98%CmgTZ|_Zh68n(t%?W8AyeOpT25U{g zwI{{+m~F;FMFJ%VcE7`K{Ed8*HV3_{06aP#UvCMkM7s! zS3T(h~L&Z66^wreb;j`NZ}D&@;Fhl@S9~qz5je1$zY)X$mBms2ykY#Z~w0S=)g5 zD@M)zDh_#t@VrDsNctuPdjNPNdeb93fzE)~-KWmBeTSDoy%&mIn(txM;trVV&}`Yq z(kqv&&r4IW@56J3HTtcy=tpqeRi(tYZ7tt@^{IcOwUHVoFJBAJsNqT<-mD$p-c!7@ zxPgiIRM-=;SrzxBbYKV(2xaOUAI%Fe_dVx>r zk6(zNIVw2iXL>y^!iU?_mEiw9oq-5uq~1buhj=s2Ed}xosfONP zUp(RgK%QK(AogpiUm>5_H@-E3)q#u$0yCi|V?27eZN&AUUV0ua%OT?Y27sa}4d>0?-t$7vU$g z@K+N4_(CSiY|N~mw|e_YD$U3u#%+}Zxu||WTrTEk9jaGa7;)c|wUU0@ye<;rf(4H1 zzm9&0D3LF!zw<*U3Ys)^4BNsmUIq^Q+NC~rZ}FojPYMFMv}W%nNKWJR^1&yIJK9tp z?WajkQ_`g$$scGlmp)IF85Tuls>+aB*|1e%MTsly--noZ9LNV#Oc!lAvC&Ntk@$Hu z6I5H@3EIK)%d?Do@FWohu)l(qlH-VEeF z{;I{dxj;(zx~y=payrr>_PY%dP;A`2uh>RRCh#TI$>U%g+tk$0r_H`A`{h*MjEB6F zSfuHBJK(Cs-(sLQYk@Uf;?sk`*8QwPKP?kB@WSVp)YvK4i8$$%yH9V&LF$5o*0(`|(v#SjBwA zrEy&?Hb1tpGqBMe__>r+ugv*D(QFu(nMeE5Bdgvn{bvLn0#IG{>Z*3>0Kd!lsRYeb zoJ1QX_IEkgms{(1@9IYpNa>{1ebjEvM>3igP$eYb^;ph^^RYoV*qw}@*3{})xN2C- zqZ`wL*xHk*g&mtGpq-|m2`V>uKr%D@%|fS~pqj)!fi;DPpW1LpP6dDM9lNnJ4DwR@ zz3puZw1X%WBddxcfw7@yybCh~ccAb4rUh}%E)y`*t$O5kXF;!Cf5+j_2g&u>3dZ@2 zVhVFS1}~9|Lo~yt@Qhk4+Efa`bhnzaASn8Hi3x{cd$cIyi6M?t>3|#$MO7cx$C)~Y zo0xRA`SbQuqDjC(@TBr@!=h`r<$j4-FbnE_d%X2++*pB_4+g3+J5mDoM~(rEo_Gp~ zS)|D*ud{bh-LD&Lf5#^tAc064!r0$gr0BfS(t{y$_Y=%s?7ubTdRE}Y`9c)_%{_y% zkTZVeE0}RCJMq4lr%&tQ1#gsW=~!Wac!O62`W?TO!2XV5a{xPbb{EHsvZ^6R{Q)6% zDZf#ks!o2$;$d%YMW0byzc&RS>vvc&;iwQDS}5do|H`p`#Fj*xUi{BENYRz$kC44% z(bDy%%EMu!$n-}i6ly~-8(KM7H+qDld%oM6Z|4$lVi!IcAFa#qY{W`{2K%9NHl3bA zJL{8>D!@oj1ykT1O=I=N^K|}9@5BEj8WlKG)CC~V}?=Vkxds9{Xg^Bl%_^`1~ z>>nz64Bt=G;I|qv)=2dI!*MeZcsl2P@i+mwTx-J4u?ed^ZR}&ql?{n7fj%UZ%uZh% zkQtlSZlG=+kLA{Uqo)wupd*Y=jcua%Haoa@m|<<~BG8ZVNe2}J=JnJ=@V5IBwdV$! zGs6Kcn@U2`HrDrs_wAA|{4{BCpOmw~4zm9OB~u1a*nZbSj|ZSOSd9=oUa;%rAX?*- z6{cqIh!#+AC0uybGG|R+v+;Uc_6A|o_dM67Z=dTfu@JQ_VM>Gd1{_cPzenxoU{f(YiF{Jvpasv(sC9SSkD=5K?u@AYaARd7Ed-{KnkR&_l75W6i-dT3vZ$ZZf> z4IVk*?Q1z&Z6vPLUpavRYxBZi`Vlby54!@Ie*&6RXvKh z>O&IS6;FCTkuLnKC?sNAQ%KJBd!m{cL0PA`?(zn$MYN`b?&86bFpjs=PuzGUX@{OZ zjjQN4T^<(KFRI&8YMa*Mu$K=B?b5&khE7zFJ3)v|3j0(?avGmQDGV1T3j&`$?!eA4u>OaX(t2)D{DKUrrj^*9M*cGz4L=jyQ}{Gw$*zMokF41Eu{48a<= z+q-;S$fkr^sn}}cB_8vc$a^N4Cq5|5FNbd10y4BNP{XnCXmjb^dj`!A z3@=qc_8U63bA(C>q(68M1QqqIu08=nKAF!v?v62B%72D#b3d1+4cmHHv=~4fZqbZ2 z+Kcy`4e?k$_t{^-*R&SROe#P2&l-L^LcF;b+Mc|^E&~xIU(LP!f%7r&L zfblf>K_h0<((~*y`nIAYf=L8~XHk+5NNw-?amE}TLN-W(~H0;yUl|bw(|09YL23>#sdJRzFJeQe%gG$UEUQ;y!fi8B}hMV6#_wI5T(>_1P>(z zk9E{C0g#!FFnmv|Ts6~KL|`@gh3dhy-TSD5Q1GZ_6i}wldpy(Ig-?Mp8%_y=2`yKf zO695$y+>5~#O*WQiR=26ftl>mp)N`YLpTy?i1e68kRs!-H6vuDP<9#?oCn`GoJx2iE6uJ@NdAMiR8 zTGYJp!o}L0?}8YI@JVTd7rSTVS>yhD0VZ3Hq#-9~XJgR-b=BXMc@lLxSU$hBNBZWV zyjcbdW$C?cz;PDoe%fR2QzOlws-6fGo1Z+1+m!LzEcPjCsDf%C%@&v--cBL5`?>}^d6sH0>r{GDykEzg5EK-&7%onc?099abgmRcv> zG8oi_d060Ks#5!a> zcE)GlrhsLU;dhb9fuY-x$?fm4A?MGjBF?oOoQiV*=B;$c5B)SYY!YV?h~#j!9yQjR z^IPjLgQ;RH5VC%Xqj~c3)5@`s(UvdR!hEs`MbAaL2{~nb2Sxf8rm!hqV#&>7OMHUK zQq6dK=j%NHBzzDV%<8?`b>UvokQTDT^`$a(c1qr_kfdJNbrSvAsOKogfs*v_Tb<7S zcf6n&w^odA8s97DR!`z0|YE3H2c%UH#B+J)ZczAetOA#;hn`fB5aM$Tft4ZsN&{TTiZx8Zf~530BBB%7iP+$*s)L*~^lIpu_4 zs(f^Okf-6KD%4n9)6$nW4zhG5)m}s=hU^be^B53G>)^*YesUD4BC%-jGf!I!o>&g2 zim5>_jS`MCi&dxC!Vxdf79&eMH_aO`VWBK1$I~#0)?Tvkns(31{br{haSetvG}VZi zlTlEFM$tn+j+ip!QnmA6;#Scvc@=xJL~OLIBnMnb`-fXr|6zNpafzZ=(%& zIgRq;F~@>s&bH$3%0g(DV{&XW-A}H;-S7fhCqAs}(K5y{-Z#@vet5s!@AbaAw;UOv z3gjI1zN-BG-1cOktXtltw=Oj-v)s-BQLXY{*YXS?^6BH|*BQwjhvl-^H^q4d&$WMZ z(R98nHB*t;<2xTYg(&li+8XaSsm*nkj@u5Sg__}<* zTNPT@?5b2PxFkqe{dgzvw4q`@(mcoDPJP1re(qw2m!Rr>@ZinG$n-2&x77swcn>RQppEYO=8qSOL$8@5v zziMjT6CAc6tAARoWZZ=#@o5Vv5wOO`4pS#qGH%o7*}v5r>=#?T!@TcAi1U5`h0!v?ba|%!EJn$^^ z7J0*X(v<$%)FHX*Q9^d0wNR_R;{84NAsIWee!Y>G6~5$se=}udA2if9Z>;0(7tw-T z$ita@=N@PN_z4d=LComm=~B)hDTTzgE6aq+P$*yxL&0Cw+b*Wg^ z>9H@JL4U7w>~9;QQTgBeJ6^c`gG`V6`i7ba3s%rr`YZkR2saszh#S415TDlCfFkv8 z9|})>OCP#N?`l7Rs4;n=rlzxJpSk0&cJX{6Ml*zY8|h+nRTWMBAwQc5-Oi2d$150# zARQ7QihWXO^{$$gF|;aL)FYYL7*N}SE4?#(%4e^t6^F>}8g(^l|3Bo5ie*$1+F zw49DPfni?i)qN%y8MRWiuzVGtR9_;UYU>k1o9WJXsTIw%};_28D-%P2rM}>`m~i(_zcw#pHGHA0!bT9~o7c^u> z$>43a(@6-_Y}z^&G2C$_N$BteA1mQS5LjM3j?stT6T1J8e|Kz&d0dLq`;6?W0qUsn zmVXsFs2IC*0Aj*_$Ciei$d6YQb!c6=PP22zv6$AQ-e$y>tIvB^NgLfKcVT-4>bLqm zk^Sw%W{Q%qeCtCboR~P|cy$q$N|KmJ>SZU#Y62jtm zcbDq?MR=+7rlG`1&8qO?S3u=#|8pi@zPCg0y?RJht#sAV9Y#)7y583t*fFdff7X5Q zhI52pH=n~Nsw%#d&-ubSE2z_+WO{&hUm$xz_tB!qWoAh~It2%L>wM*2dV1C#mjUY} zvzmn2N*40_cp@u1BHD&pTUi%Aqodv|@Pi49;6q>=%d6DSx!+D6nj)vo~*H77Y z=|jKy-0uPs+|EsF>ec0Hl4fxKtke}L8p?y*ws1;c9%wuA zU>(rPil*0kzGreF(+_hT8Pkc!i3r61Ja+{V@23s|87;2N5#g6C6D)893O{&>?P>P> zLe`b{&GSTGB*{Umi1F1&?Ry|m={{C@;*}3j0(!T37>u;nm4rw}Ur8u`Neq8WHC#`h z>-~3*+W4pQ+BuBVFfD+_e}R=tSI(+KIF0}m#Mm))L%leZ`zC7blt<(CMAaPW!R1yh zx?RsjWelB2KM-_0qk8wUz0p8EkS$2*^!g=&r{}}j&0josdO;Jo8oAz6d%0*z2LI#+ za!KpR`8MS@5bi0Yvy5}nd)ZG6@qMGFW-1}uTIb(c=vXWNSP!VkK-l!ouYga@w*`al z{iJKx-@5<-NiXK+tu4VH7B{W{!whmiofsRW@L&pZL{G-jt*^l}_>ZxIIWD;P0n|X- zOF(fPpRtB|Z{H(PL*OBG2uL_5q}&Bcz7t!z&{CLooyrm{*vi+l-mIFEsTt3me^1v1 zCY7RuKINVNbdno#aN|F)$dvLCk!EShOD04d zu9iMo2I@!)1;sBO1kJOL+?tWVnXYEb5jIZ4kWupQ%d0u2`_4!5*n213t%NdGFJvS{ z*l6(P8Gd!=JFHrPcF_61&QU4g3=IMvHI{uG@9yok1l6V%02czb3CQ`!(HeMs-J-S}=zS4PgPYiB>?un%NRY>VdXfmdY$hQp zvzZjv_zS|+V1ae7$=TSDp9v4WyM+(LLVJ+Ku}{nM^g5NiN9SIDXwvHiz4H)+P?Z8i zlI`16=vPPT#X;vv8W+=|KNnctC? z*1?9^AcBH@wU6WuE1{-BaXCu~@>i+itsH-sSGcNU=h@$87vB7=PT4A=MQxr!zxQyY z=S^0*Q$p2W)%#c|oe_Csz8I#z9d*Y@H*Zw*TszILt=y#UFn_#X52VANAo;+OT#dGMjE@}zg~M+jw4dwf*IPMQf7Q{m7Es;pn%8TIP;E>ldv{NCAhnJ|b3km$GR{ssrfI3?Nz zIlc0{7cwTD%}3&tN3)mD`W3e5Q+sZ*3tTjN?Gb|Q)QR(#A`vH0nW*Nt#*U}m3a4*?V6!SKS_9a(mEiZ2Tnt*G{L-cDgqBbc3 z+1!G!MFdhMXw`ly!W~Lb2gz}+71I-6f%yn;gAe$QZ&&V7d)`?ux`Ub@b!+LW@il(7 zZ%}3N;LPezS>Xl6<#~L&#O4o05TobfO0T$Iv_(WT-O>!PYVe;2bU5zew^u$Ve_!`P z&#_lCFBe3`?Tv=7d(z*yOld2>gg@T>T1lb~RUVnQ?g|ZIfGTD^E80*0te`8+;NIyF zogt8NfZK#w9QFhEpsk^!{ikyFUe0%yr^}Tz*eZQ%hA}u^&|^_Gu>0N3y??eXhCZqy zelD%!IxZ%pG;wE@KN4G_sA%xI_p7>*?|Frz0+noOJgvW<4+@{mem?#X+`U8as-UV9 zmvW^D4zv2pKxjnwZ00uM?Mai2`6EfFs2_t!_%ILqnXNsD49ShUsBRFJF>HM2d?Y*i zEX2SyytQ*aQ#OPS{c<}|-u~;AWw|bL^4z7Z(mr&#^e0Xfe|_;&bb|mKg;bQZx778V z?iB-b7x?NYbcEmqj$YJ-pNoxC4;e=F_o5zNqTRq|=)S=R2+&38K$K7T^|(Cgp6PdZ zU}&S4fcKei7@CdO!Jd2#jtlcv>_c?PlYz(sk;J3s<;wshbFYOZ(Y3MAY~-Igp7kfi z$EU>&L2%l%Ds)fSbL8hlAQ`r)kzIxXJ$FL@z?gnBHU$JKgI71i81Uzl!H18z`wUPq z0m>h7OHRAKBS0^eo@=ar%@>?XDs~OR7Cn%6VDbiLq^Lr4`XrLhVTH?}7DUZbkb(SI zU8+QjQ@?1^=X@1A<;El>6D~geWFY*d6}}r;=8Xi;8w%r~ioy~p)H3ZSi`sWO^I&c8 z3#mn2ahbb3jESg)X;mSU2zimlV=}FVdPlxR_>M2;EW&t^Y-!`{3u_|R>w0l0)7P|$ z%$5bFHBY)I-+h)e#FsE*BdFd|$$_&+O3djz!VzMkle5z4!ru|$gzyA~?~#!EdN38u zDAe0M5^$V-aFX1>3j>9z-XSt@rBM8jdgw9#>alR$n-@SQAjK+hER&3I|0W7^XC(I* zd&Pc6KQkyQhH#`QB1owTK}2@^rO&1}fh2UV_fst%!!$vhjq{l9g>On4gWRm^DkryV z+YH7&Xy(yp{teSl2#&)2nCVyuJN|iHUR)?#GvN#i+3P;@Y6M^U=RJOC6BfIdD3=a; zT-{jMzRMoPq>>-wGNX z`pI7Je+%@XWU+4%9-`6+(~sTtoiI9 z5d3p*53D>jG+zD~A-nDa!D}6*GDGw4c>? zV$n|vs0;i)=boU`=-!hv5q)XgsP`M%29L!Z1mAxS5=ccxVlxn;^pU!MAtLsimm=M2 zP3I9;X7kyKwKMgy1W$gmXGib9)&cgFrg0IeckI2b?liHUrfPR-yh-A`k)p#d0KFIWvPxi;yho$(`E5B;;Hvk$Oh0M;sg9GfDO zQd_+BoLR6*-w6x4usaO+2~6ULAD|Qd+k!*PQsYKD!5Alj!1m#hp?3g{QakX&RSCk} z*ta85uhEhT1l#qYo#oZ#rs&S?%qJn%WbM>{DRMSpRjo*Q(|aj{wuCZLWJu)I@=oP+N^zli?gW?OT`VXr$@2A)=0_Bp{wR-s=4 z{{F!X6~xeM&^mx?&wQt1qG8W7`+DvcevL{z^yIa_7+HLV4IubCzk+Z32q>&Evp*zF z9gVR5n_qBO64q1iVd7zKcS}jLX7*CF#={L>x9n@4O|?ut8zVI6KWWeEw2P?TWb|nn z{PX_(0{{(lgJU5x0m`QMppzI7i+fYeHXyZujHa#es@MnDnTZ_r#IG*NAfiojP1}j zmb0B?i~a*z&rN(pfi)0e9LEcZq~|QuDP`AG@5lE4JC}F;cVqB?vIc(j|B;!}?cT$ke^r2|gr(kvyxZq@8ltR1WHPV@vz9&=~nEUV=B8d!^-ar~;n!Yva zS{xO~5bOFL1^qWeW)A-pXs?=Vh=szxTCE6q&^U)APUv z1M@e-30l;CyFac8Qx}%X{4~dgqO+Mq)<0g{vJcciudTP4RP2+*-dM=WrjM!tV3it7 zWP-vHHZCi?kDXH@Srn;T&w<|R3}YtKAWV=2kNT&;=0PF~zyL)E_n`Ji6jg2=8+qH{ zvO%;D($Tm+1P~IR^LWYNt+RsZd5%60J-Jds^w%xso{A5LehRI1!X0p*;+nO{LvQ zrO%LzPH8CKTZgO*69QqR$LeK*yB6*h*M(Lq<=E{V`xx#=YF1?ZT>5w=2}LO89$<6hXsM}n>z>u!nH&;2!O3rN>!d+oBNYAO0` zv*0N3^+7w~O@XsielQl>yWbAoO+4#IUds4zYydGq3vaBIU3~WA9fVboKEkXfBydov ztcezLSpC~oR)4LU+Sw)S*$&!q&QG|q)lF5~^?rUeQ2`|gHXBV9TwfdZ{kRA4^V^l%<@PI#^9cK} zz)IHms7G&5;fwki&{J=FRRVve9EhpO z`F2wA!UlC4vwNQ;ktK^BGb5q)EW=!IKk)pq*;BVA8PECRB0Z6-;^)`0mfNc;y}%AZ z&^83epo*AaD_q)VHnDYoI&n^(d%7(8s|iEcU)(Z#p}$99bq4$G|LPd5QGU1}%zAx2 zp7wqrNXLCULb?0r16ATFjLQ4!ksMh!r069`+ecUC?$W$~)&Ce!4PcD)>a z_RVxm6-l)tC*WoeTfS*cjBV~B-#1&p)!u_PX+JCt! zU%&$nVmPW;mJ81Q^Z)H}_m4(M*54u7m`r;?tkCi|O8P83M~W#aW~r%W7{5*3 zuCne3riRY#+skisMj>Ynmz(TSr}d4Sa*u%wC&E_}$jO{4nb*}0;grA__MygnpkzLm zZ18ZCiku4?SwIc~=Pt1sHv>fwYXR|#IPRy!ZcXCz7g$!G0chn94>MmPxAF!JGB$rG z0#UnVMMBZ~1=@ItivJ<%gApLJUQ_m~&l|43>-87u8YUM+2#Cs47_| z-jT_ktIg>#8A2`bM40kc)B~Q*3p%XMxK)-v4a1p!WRx+OdIfKv@KuHV7oFJ&&JlVJ zT&ln1Pn?rxFuo(y;p5^7QRv|?jGDcl@P*y+#CGenDiAFc%cRHWxvwMC=YDZmvlbm| z(9IbPh)0_XOu_2le(Md4ppmQP&*Q4#LkYJVMzGZ#yggnPxb&hQ5USkkRY^U6Dp5WJ zB<$=V1sYzJ1~ko`Uavn}>?M63&BK-GZTSrMLWwvZyxx>JX>!W5yz&XteYnk7!n(Jz z_2B_ZnqYE6Z5Dt`$->~5so|L30Lt>sQGegF@%X;E06^NFu<*@}(YaNJ37in{O+y0=UNgWoQ>#l~~05g8jJl+IR^ z`SJRolYY}^>-J~%SK~bYK96VOH$v=Kf0Pt>g z6JRR*>tfq5FMoya>x3BOTd-@zO#d<3m9sAly#2J{t?lt;Lh?MvvMLIdek`=K+V$w*1UZ z;lF3Y4%$&iNf*55jkD9IzbTJonC5H!2;U0agZuf#!G|$vVixC9kJ=l-+L7zsvgPwV zBe2}Kdo+RA1B*dG#s)I%8`$dP9e?93s-Cq;QabJAXs+vfZQXXroh2Uk)11ZP#z8GP>N-plIY11*07@m4+PH>lHBlWVI3a^-Ypzb6SQjco~1`^)cI zIxq=n;F_v0k5~iT&#T_qm85<4OIeC9TY(*|eTf2h5er^;P%pQvyM=`IA((qLqtPRW z&T4^k1A2AQOn9P9;$z9FeyNE@Z+K_GlROnaM{_O)F8Hw37 zl!fsNO+RrHf~N7z1h^waAUB^b;|Rb;@D^nM6Cn4%GayNk?9g-zqV*9G7tAg<0`|zD zw&uF0WO<&SfB}lf=Pgfc#CRLt!2yhVKN>ehnQnm$j5i_&U4C2mqxb}%!Ly&!K15zf zMT6e|c){c)@n-VL()oBrlc-MIsdyGouQ8k=4R9|~=;C`JQRU~+*^dssD0^VZe6_;y z&<6G%mMR@SL@FQzfZUAu$P1QYl8dcVD$`h29k;e_PwV3p zbwh7mp2h<;#rZ&Lu+wp8LUPysv=UY6v>>DlBG-(?iL`VPTa9BI>Dta?#ib%Y_Muss zzFp$|sEA_c1toT_cn16M@X#q#xo@ zyG@QKaeRw4*$$BSs{d%wSKIA_D%j9egq3t7G6}NSUP)0CAliJL`8s z<#r}H0!37w)KoN&9<@~_Mt1Ky$|0!Yj7(8&Y2H{)NR`HK0m13x%=j<#z`$uGszd)l zxZ5=&?o61yE-37m7w7h3%vg8TF9fn3QkQ(ujnQ49z(Pm(alf)lu%F=EJ}Rvjn$s4Z!8|VZdlOi+w8(Ctkm24FpxNKLCbmi-4Vapa>6vbCl}o zrf>=Pmq+v{Z}ofTIf3SP^7a$`ZYfX5Q+ppr;(4TKJ&$y5Y!9VbPV&g8iRkunnj}k#ouBag{oIoqB1EH%e4f58@V^d^ zUL+f)n0dHD<9S(|f>G2<(0+q=g_H{}`KZ+C_u1!`NAVNBPgM44 zHMs^cH}Q`_d5J@i9K&^C*owSOB)5+yV7y6(2N`(|8g>^W&5h@=e{S6hh~{6`j9fh& z-Foa%0bj>>$*-$KGL;iw`LF@8{IHHOM&=e%6L5pu%}9m=N%)?)ZnRA`Ir^b>DR9*J z2-B9IM2GG9y|NZPbyvV*F6LJ@CUqZlwu(FezlV>cKKi6MQdIM^6Uz28nr2*#jcC9` z)5Exe6qvm`Q3X8*F3C*98WsEH7K~Jd$++in!TLi3JWPE9#jRMsb*25v`d{pwOOvYF zwyl57sq<7Uf&u~p${T9DR6yRKz^}iJT)TK~_1=gbE6zMS*36ljslph2^xoRn7U4)j zwq2Ews7=o($?rGqwH-a_CJ4%R@Gp5w5JV^@vc+BRjHha_*rR=Ndb^lGzLz3A6)jlk1bSVX~(qYQtRch}W`eRzD870oa!ufuD*iMt! zXUAZ~g95dTj_5~|co(_tZ4(NI6fzF*UTy(n=W6EsSU^O>zdAt~3g^Na!yRmFGJ{Jx`IJ4*H5-?r{3att2qM$=1OAtRzu&m3i<0!}X#Hk2xhYPHy&D&JVE+!f_eyZ$|Weo)Ap~S8sS9CCKyXb?M5A zRZlBIiV@=KM4K1iwJZV77bAG{)eVRuSRFr1eo406k)Dc9YPdH*1neUH0k3`iQ0(_% z4!{W0RI8%UoVL_d62MnCjLQ3k4Fzbtj?Y$OZcMjr7c23NePcuyvW6j-x$<`gHL>)z zQgN72ls>ap~q2vm=%ocx4Uq=Z+Fm@68gp6Y4;%n-O6AEH~j3>tT>{JkD#6BsBO zw!tn|f8FQh`br+m5jpc;_4;MyAGOJ={(O`7eJObdCSxa4n%P(F1Z(F2dNtG3B2K*! zFlkRl3D^Dm?3o}MCeGCXU778Id29ZJq8NpQum@F1ft1fwD?wgdYlch)YAY(ghs)2t!o*?!UM#-q$jmj@lr@Mb$W?1c! zAdr+NjYexmqG)szy9%O9B$}LIdL6t6?4d(-`o<3*nI19qeg#DN(9p$*Iv3yrvUK84 zxxb&#M;pC*r}qP|LO`bas=kbSNMTdQzYdMSBUoAgFh~45e?_0mAMthk1u~z>Fy8wI zI)GU+SIzjoH z+K)?$#!1cqndyWk>9JLWfa-S9P;L2i^!$c*C#HpAulJTO>yLr%4Pt2p=O|MBbw(6g z1_Iy%XBHqS3EPd7-|qk+B3(fJ9t5Zhyb_rBjnQ5w>>wKnp?9Cv*LNQMyrq@!ylG34 z;(D79ZUJH(a(E6|X#0bHR0i|RclQXQ;?sb*<6e95NPOnKNXJ}ZM27q!amecvnFzfX z8R=A#obhgdPra};L{EC&^Q)&Pvl)sOzpYmsgmd-0q9y9f)$fS7f#m(-OT+yd^kuEI z`DvWz?b69m0fdwRUTjG#iVlSULfc=#QN z<&55v=AFDdAX@(ACv#<-@xSuGut4NEQ_%Y#;|de~1u}eSWKQSOW2eg+lLn$@u*}Nn zZ!9zYE%|=RbD1KaA%aQ%Y|;3|WTD|uM|`cL7`&~9?%3)+Apb2hh9j2pZAL7@Q^?U< z@z^B7Qfp=$7GvBFl2`iFC3%1|q{F&ovm2ysCb$4XkG(^t$=g8BW-axD`V>!Rbly+& zc+dM|l^M&O=!(%-Z)9<)NI%b@Ld5Ao?Xb$S!-XstYcFdMq4fcg1^9LTChIu;=h)07 zxuQeJezWo9FGoOfyzdV7lskh?d(|ZDGT&DsKD#eVz{hV$;Lcq3vA&>ZAJ8myPChKpF@~g)fO@Ht#V(hRKFwX ziZ*`UR9|fEwfrcv`EEfy&jDwUPdy3PRi@Te2L=zy`&>=f?DGAMA}dMAd{dOjlP{2R zA+f9NdtQ~&rFmyxNI~z-UIWb}WsvukrO(5Y!rh(f-u~rZQ-6(*c$({zVQ1{e7AO(z zE(Sn@FdMUetDAzWEw=dS8~<^y82Q9c^w;~1Rx^4%hrjNx<<@vM1N}Z*8aJ63FxTK0 z3U2|L?PpD<=-EePjg8pFPwMcm28!)#Lg5O-dfo*`YyN~EoA;L2t@#2gJ;d|6?vXP4 zn|_%O5DFEEn7X$>=Wg2g{`IHtf$e5KZu}&oRUK~ zC<#x>+}@DlwkJc9CeFjTf4b1*alKDo>>qw-oa>LINNtTUnX%k>p{rssShwuKys0FY>szyG>|8W$LXghIf&eB8r*7M zu6g%7MZo?+7rT6588fWSI94b;nMuZoSi$9s^L6)mnTW_6EZI48*D-EsgvQ?9^>sRu zc9aIs@a^BAq`zIKr$&lI=cVV$W)-F%^qYtMdDs1f&wq5`sn-z@P(M84a>t?i#%bQng#H(b`K%?EV zY$%{WL8tmXujb`v*{~(tF5YmpS@ZS@34?Oc=G~XmgTA+2T=K7ZVkm3>Kpb$iQYIBM zkxO6GQ}HfBRIu-(h$;7X+qAtqA4(|x3+|L&VTs4TXS=1!Z zZ@;Zg?^i|t(=0%ujzpmQEs~vHdO0dvoW5VL`L1|X(tse6$HQvKLHlN!-rgS5q#j-7 zJBvu$GQ{qEUFQdIpfT{n1y+24`+i-0bU)K={0Iuxh$5s4s=wFQd~&NYvPyuZp3o5} zynI0xUYR<#Sv3V9GLdCSqm?nvJ~uKbPTqjr0NE3()p$&>LJ*knvORg)jdcP@LVZk) zk;0~{1-)d77%!?Qh|<_yHeprY+08OS4fxo}TLWvnD7{RozIDsC@iO=&GR_$ExAul~ zjE?jH8EBuW{r;rraIh|MCAE8(^2&L)w|HOQ?zI&H9RVtob}$JJR^$eQ`eOYQZ|j;) z;`x5+?ET4R?GPV1!#7xFnhej#<>czJvxC6zEKQ}<3mn;WvnY#Lesym}T(mZg%s4TVf$`rmcL!N{XwE>zJNkfv zV}fcncKmr#*^I`-wilB7R(bpGw+9BBam0vY8?EYx`(4=g3`hNQHf|ModjuL>^uBF} zFA>mtv~Q&3%80m^ug^k2qDl}Kr@ESl$LYbYE54C2V$29R7$zn}F8IYp;4$9!Y&1__ zctfpTxmY#aJyR20+reW1BFc_!4Tt(?Lso=9l?j0ZM0{m_#24Ak{@>ibvI|Nu|>C7kW&81emQ5unv7NgG@tao$r^7dR1>}5%l!VQL=m4f1Tez zJs(dI`QXR6$bYlWX>dEDy%VH}XGHN^m0EsVS)||H| z7>cqYutd_$dDD6#$`Yldoyfg5>!i|%rKN%#h?j}8l)ms^eO2#1uA ztyM=e@DXp8wLy?_K8x25&zX6fLLy1nL$c^TickKIs3eB8KmOe{pFY<;<6LjYuh1D? z?RHiKI`Xvs_+BE=?M*#@_M7;AHej1Tm2E~gFb8aQqtzfqO2mo8lg5Z$M09|WxUJ2e zub3Wgm>C+2DcN*U*{KJ#m;7#Sts)!Pj$k!yO|0G+G>}_GW!j<~%;@>$M}1zPsGw zq_l}I<(xsO^Hj#XywDn#?%#m8GZkUpf=r&1bFjZd@pmVMN?G`a#B74ke)hc_zpim6 zsNd}4uI)Htggc$3FR3jDl%fnx|A_PVbheXuQ}O**^Rqh+1T@%@b?6hCUS7%QjdZTz zTli{A%U%dKEE(Q$VM&khexaL%;;M4c;Cjr)3#`GFiA+jzaJ*0NpnfPgAGGx+rmH=_ zynZr}ZOEQDu*0XGR#rQn0O2_8jcSE}Brrl8;4R9LK`-J6un7p7D$+i`cgWRcU;?X} z*|DC8_r94vbV~bbBs7aU8)l5*A5g%<14=;lFkO$vp``I=*xncEit`RKj`7{`o8K{(j!TL-R`qsOhGo|K{gB*oy+w<+-?|K{%urvMSrOJc{I0)=InLnj|K8Qb z@nHpN)BPeB;BnSBJA@KM*cOC!t1d2T&huaKCI%4Bt^Y>^e^*=NUGw{6;%>BdBN|4+LKfF;c z=aD(Y`h5G&9+)madP5m=f@m?<(QXX^BLQU3Wgk6;11>2YFCR(To-equk;Il^klR*9 zshF0#MikS_g+EAVbOBe`fev`c;O~>xelb z6imVN=S<#{)Hxjw8JZ?}kRB*w^j1MzIXn#gPWuDgMh(3C@)YM`U#Lr8|11jSB9`Si z(;@C!!OJd;H(&R_SvoO6m7BRr6h*DDQjLZW**?4hoR5c!~^ zHOjsBO8xX{f*C!C^X4E_?;Apa=1nS&LLr5FuWPv20bt&L_g`-9b25Ss`CC&gW{YfJ z;ZhaaZPmWVak*`DS0R839 zU?bIgJ-!fViB~VTX?V*HZCnbskuGHc#VZrNk>{epd>zL}1dUS1*-Xjp$IjtJ_1)1J zGRyBU;l5&1m=3posTJ(92@*T6t>s(VW|;mN(H9>D$yRg!J6;JD!S?Cl$-GN|3D?`N zWZUcAe1_1G_^Nij;=qg0$g0`cwu4r6M6f$D;6#rOkSW+5$Zl1Ir?$TazWtq9YQ?OO zesTyM`EWIWMprdGyXs#jF23e`Gbp!^B@s|ooSoS0{6FjJ~_d_sG{e9lw8zbM}6eZ6GO!rR2 zk}L-KcFSG)CdK=kp?Ed8KR^C6BzsY;8c}goE`D_bDwEh3t@ggRvNfc>o0pWnihIj| zdE*}6v!Ol6G8cnV)adhi(1>BTfok0!RmVtSLNr`CgQ!XZ2ZOIk3nFOWeT3x zZ~3{2wx3PfC0{SqeyxGb!|_;$mp;zz`~e0o2%5kyokYG3>}TsQo5FzgRBrw<6qiLpx~Uz^`gXAiRNPDMBA-}QPEJ$^Z}Jwhiv zKRoG>(39hS`yBPgXAL4vLx5b>zta+LBFNruZ`1?&2z0jG!>W2VSQLg0l7M%K-uDp+ z_sGQ-u~1kf0%rbCC8^bImnDZtgE!%!AJ0f zbo4?UJaSc~;TVzp4ZS`VmmA-ofY$xd*`W-?0;8Jv17>UrB<*Cw>Kb`eQ=v(EL1?wk zJN}d|J+r+;ubIJ#228!w4l?!lBFCHfEl|k*`p2=PJHL38Y~3GU?&bC_-6~<>A2d#N zp{m=12(27BM?;(akZT;D52TnklKzWtLO8Zw(vcl?mam4luZ`YFCMs~Lz<&YI5(OSQ zoX^_6B#G6W2ZG+O+VgimE?lqW_jqSNL$vlJWlc~v6<0h1I-dam;+FhXFY)1Vk@)X^ zgnAO{{;#VE66)GBKIYmG>?2o(?mOq8)Jrg1>fI#93Yidk3T(crc7)yQU<0R!5Vljf zR?!qQOGkbuiuAIAA09BC6x7KJ@yE( z_opgKBBztv8s-Q*L*J3x1)na5T`mi-CxGB^`s5OPzl4j!m9mP#-OcBig@XBpjDnS| zVmqSNP2`Z!ZhG(OaQ?lh10shQVm*2E^wqIHA_UW6I2w?D%xyd4pe$e?to8f^N*ofE zRFeVkil^>fq@^PCoq^d7%Y8eq7VP3v-$5y`r};g?WaH^t{yuuvJfsZ0a0H(G^pS-l zw42dX%Uus2i1i4v2FGQ%t^=9Pwb9lL%Ek+mQD^`1xIf`wdaU6^Q1*6FK_H%#XQY4) zGvaxV`-@dw%g!HfQrE^~d82zP(woY?@u&mp3G_7yqudW3?I^u}|D0i&r!T|Q<7``2 z2xCdqz0x{W{~zeL{|%m5G_0t9J+l`Cz5GkV{g;OOuV;2&OI!ZJ>jKX!n3#w>Xt)q$ zNFL=k0-zn0&yPCENQ3@++oRirIA0j_#Pn?HZ$kcw#!3(%dqwLO5m%_jrK2_hg|5pJ2i zMi4XW5pkdIidM~vG2UJi43(%ie2=QweG%vrNbf{=CtP0{5D4Q?L!ouFr9*UAm~Yt9 zb{WS@ME#Z_nD>L3eeR@4?3--Ws8XgDr$ZCoe8gj(R30yNC3yYFJdVJCFA$3-SrB@O z(b+-74akYjLq%3uaEl4OKF6)~eW6H0Hl&2%<%d)O8pTiqKjrswWnMwE;%yl4a{z3+ zjZ;~F?jKlIWC3YHIu!CJS?IM4Y@jJlz8Cj!a4?b^B%E#}pILcee($Pswd+DEaE=a{ zz+^S_0!lb+t5Z8$jd^x)?u(8d@;I6#9q&bi=Ch+&j_jVmG&OYec&aHVxb8%LJ(~B=)!sj<-&srxo#I~JQ#l642WIXgN_go#brc^#(fsV_*Fn}b z0DobqKRmCOlD|svndjXek#@fPpr-B`TY~FeBu69|0^Jy(V`4=#)czbF%86W;{$~La z7r#RO90mLp2b9!vKTtB;+WZL2@2|bQqM^<_bRqP9ge0;Wjw3^atj@4NBh_qCFYdfOSJl`wh4xKRDW|4u%>zCVp#ijw2p$7mx4|Db>4#e(>4 zp*l+h8W{MiqbHL1pIiDkArlh19*~3K0>f3CX=FJcy&!R%rt>u077u9R9%!VA1K^!{VqlO}ib?;zaM)iNF57ZBaD#a;Vj$ilvxexc-&7X!t>y_K;$F`370LQ$vh{JvoNMa;`L3P}q8Ewu(asp^6K2MV9Yyt91)q1__|Kq$4m&OqNCxehb4uaef>hwnzbB(eM~ybZZY;nY{5Fv`5uuXa7==mgO!hy zPXpx212CZcI(8`0`|9(n)-(9j(B-?v^;eV*#Z&>@1W^-caAon)My9R3*rJFT=0DYQ zwxXp|C_L+ayA*Rk6<^#`8(TZ@86}X zC0Pp$h5c!`7|-pS-Lqdu=lp3m-04u2)SVjp5HK+VA|2;03uc5XtscAUs-Q+|p0JbP z{>NP+h|C@nSj~-mwh-gT#J%SSE=dtO#0=^^%C$>!EGHV%6^k@wn}Myh;)?GL)DeK& zJlT8tUvJ1+-==8e;5w6;<7a<;^=`VebL6xMz6A79Aa2m@- z1yWG;u4fL|a(E@9`}wm1V{jJyvID5%^ot+iK#CNS)ByN=i+#pxxJid(%phjCJ%439 z$Ni%g%}yak~hUl{;$CG80*uAHJxH)x8i{?w4NN5->~Y&POQ5reG`)= z)YK2*=sBeIyI!=|)AH#p9k%asfe2B;=>J{_q}DGh{9W(R4wXA>S8WCKS%);H)-kn@ zyxmh7~5b5S5?jA2c)E>GnZ++o@=Q#$(&AYZtaaV7&n^4UI!3+3d zKmS}x5!H<1OE7?$8I^Mij!K`a<)fp@IbU9f;3kL@oeq+?i6BoV`=AI490>}GZg;{0N`<&4~d zHFggP%vi@OT*N=)D$3732G8 z!(o+Fdc<22=h)`MaG-kib@6T+g%0EQZfaDj>YQiL)NQO{Y~cpO-7EKabef{_A0`0W zjtsubSdLjhVJrKz+`C{+9)t1)d1%E*#CK5Uzj?rAWmrYOMar%H%DMp)`#XMSWRbWE zsgcM5vD5HSKl^%;Pj#r}r-z^Yi0HtP&il~@dO7^{&hZ<{f?yl5=HRzr=&#k zG2Tvt8L|jAwomT}fB^`IIrm`O?-lW3eN5)7e588^nV8P~-Q<=asLGE4>jlpey;68M z_s^=HZQsOLi#HN4b~3EC?#Vf*6|gYl2Zu>0^4NtOhMRXH?CDTo?{*@V&3v1&R%_nK z;UIB$T4;z>nMjEF+|*bd47{jgOQ@rHzIaB6XhKDT>7@T^ztSCXJbMB*93T(w8ZRu( zP@El1h3RqR75c0zhdetuUYCf$QB52Jm_O_WI}K8q62mF z7h?XSPLAIOCILo}>+Jp5=-vKhq%%hb{4fO|zhocuFn=&Ww}?hXOCT4%8oL zzB|ZkIqg!U->F<{d*br zGV9c>uH(;rbTzFPPo5}m$h}5ZM{7;fM~X+n%I}$<+(e{EQZ9QsUxsWC_9Q%kj4^@) z#-qy*;VjEI6z@MTOvdHAQ3EP43UYZE!(d zgU>mk!>TxN#H0IWuN6$&5lE7k5g2fG+AEYd_#U&J!uiVm;Km(;?TNT{KqEBQ0DM=U zPU5VvucHk#@lYGYn|;cH|9coy+A-btI^pEW48Qnf;4Bgm^Pu^I2@xxrR7|1{8_RH` zA5TXQuf1?zQ{}zl_iO`nO1BG`w?1oEmAitd{0YM50-sS%d39RVL$MVK4z)2p^Rcku zy)7#mwu{5(ybo^TnHHAlc-WZam;tF#j7o1m^_RwJ*VR7bXdj_jB5N31jjcVoS4Z7^ zPW#a5r2(yFz;a5f0=Tak{5ZX@-4nR^_WJ@#vjWiYXDD4jGT)>ej(YH0J;;>v zRb`V4qi+MIbJ=h4aQDjy~9bN%R#3o+cM z1lgUah{mr%J?O#qtd1K*$}@H^TbBI>exIWn!GmV5R9MgPe)FV23dk8F0x3o^LwUct zk^hT+CYe=@KSG2R4|~#fCp#$5;e*ryeRKE69p^(9l?~ZJ;Nw&3z4MO#@p`;xvoG_A zxXBmX%?Fo>JA`jE`?MZbxL)9qiTlc@KyscUVglUq#tS7{0MjGtgXMav-)P+m0etY{ z2d4s3VM{g9&uChvkW*e4x(^?&OZ>Fa`yCyO?&bLF46N(#Wx6)}7F94Ru0dVdL==2T zPI?C!M!w#$5=nKnPZ8zAFEuQ^4oqz@-#{eTu~65(Z|Qyf(l1C>#R1_1sII*O*M$QK zG6**mcr%8k8Ao45EHZK`)hC)ED8qjTsuSyMn$6}re0kXQ=d!g`ThbRGV@YnxrJV5~=!|SIb76mGqZ#^Z* z6X;WUyu&7YAZfy#gNJo0wD|p%F4S>fdp{1cpiSnd1nNnHM(a5qb@I@>=(6U|r|?qN z^SxFJh{cc1GOtrUW0C%mKZD3M|Jp;Ry+BPryxJH*2O0_=eDaUVyO`;>o@*Q=tIG*- z6q|*zZCqrjEZZ3VEp86dix1R<#gX2guGMa8tJ_JRTsbxry?bil9`PT^dPei+UL#>_ zBzy{<2HE0q7q+w>>Td-JDqIHp%uoNZ)(Eh?N9&-vw~s~bPP2oS2`E{gQK3TzLjyZr zlhr|tn7hqKY@!+*DHA|=(NZiW^$}y{>em)%MTk*|F-c#rx{xxBM&`gH1EDh~U7BO! z^X9MiYvC`QpETiFrkr@DhSP8k5?On{9}$>}KE0ZkVw|I?d{k%^L4=4NaC0khiS^6p z{Js@!n7Er=y>rd8xLzBbIEoqbh%}$cPen}|<7cc@Ayf(EEmN49Hi|j>eGEE+V!KM6 z-h2l{XRbzp-5Qig>xpvTJM2A7#1+Z&xhL(0oU!B^?Lf4%u22?k3CRnGtQm;UuLtaI zUmDnJMBT)uD&`v+b_mlj7#?9~a(fp)OtLu>;Y6}YK5xs%PZ5A_=IhkLUI7T!@vb)T zA$6uPh!~!)b8pH<h?fuYvE*fEamAWqoWeQfkO%dk|G?2Z6f^ZcFe2TtfrQgD)B$)$y zK>1yF^`=1-Lq_S#^l;dhRWb`#KYj(pFTfn3i19?$!_o9#nLun`4u@!}ZjAVlS%CD*|s6`jxWE|n;$K&fY z1YTIZ7a;Q?e{AY{MsQ-KhY$GeO)EV&(lmKpmn}?ByoArdV$+ ztS^po`t6I>9r|L-NI68*aYKz$&2o7SRE7=!UhPUxFj5fAllQ5D)_cO382$9T!Tv1%ypI3wJboQ3yyM2CCo*_*YX?U5#wgi;hVe*-{Vd zy|3f%44{$dMVuo^oHG7KygX-d$msNyq7jQ-J%L3`TPg4d%3AJPL&-dheFj19%j*L! z@*vlfFDWy4_g?YW5oi+-1`~;K-)9s)zLi*lxS`U-L&>v!W=*2uJC}v~|F9mWiBU@g zLMF8t>y2VOD0ow4{*eogI)8T=vs8}vUoMDH)BM>^a0fKZty_e{|(nft^0vnj@Av5dkM)o zw}fO9yrljo%DKkx-gY`Nc5Z(OV1CO#8p+tuo|Byif2hFhM82VR#e0>Y;C9qLYz!Sa z-bfArZ=-&bew|`Ak zUIhNoz&u4VZZ3bP-TKh(P0g>fPu!tqzSm{u@=v;@5mO2pP3*T13L}^Y&(tl2NUe(4 zs`@2h9w2XWpLg*$j>G22{(_@oPAPq7w%b*775;=MKk^+CLeI(s7#WidD&*^TUQVWV zyg4T*dSB`cIyQcAczf`#{53gAO@F4HRe|v0U3q~CK7aG(Uzm~#sE}w*9X&`!S}oUj zO`O_uj&@aoiN%mZ;w;-dvV8dKUDQhtacV%AIE#14i4$t)A%#JZC*1D?$-i=Q$qozJ zZU9f#LyC@k%o|!XlEGeQ87`9j`pT*Bp5m|PT<|c*=s#DqxT3PCtGRA_3tOwRzaXtn0<} zkU;)@CtDt$lLVj=m^dABxCq)sG;cAfQS^8P=tiTFB%M!VSnFzfhm31>YtYJ*S9{?4 z{KU%(^CV`QY;f>-vKHZb6xrQY-ob@}1IE`eEOfmBz$Ctb@9)4|oN@7Q$PNGWqPF=1 zN;nJT0ovsxAUFw~YrV8$P3@;j8toY6yXVha!ooa+KK1WtqI4WA(63+T&8lm5zta#7 z+kbD#dnSDg?MYsiQffvAJ#vYhnwwb?gaIn}_x2DzH_c0=pj9nD)8PznmV1enl{{#(jP)Ud6R;02hN0nfDdvRR22{u3|FiG( z;BRh^0K#8NYFN#l#RrRshC8)AkozgQ=%ouBia0WY#m1QN?=JSjS8I8;2fZ@t((ZTn z=V?s5_Sfp~;Wo0>+vWN5rHIKkJ#G-##nXIB$EEIgG1F8p%ik+{LnCpl4cV8J zW(QtBZz<GaI#%L4oB)fucra;a0{UTH`_Qpc=s$&F7MsbCaK}uqhbE6>*}(BMurY?tk*BbdNlKl}Q1}$lZRO zQv*bmKAwz}!5(xZar|>SnT0GB*QbJ9HVe=ycUlUt2^be{1xoqhfIAJqYQm1cOTU6* z;#eu zu}nJ;oVe^Q#{Nk0phI37;1(tt0$2v|>_lf6z>?p?j|0Zw2r9k+ji9XIgoO-+Pla#G zs=JS0L)y(LT=;pzCWb7v<_F)D_O-V2>+&`p=+VppM!jgl^&0ztUaT9a4;DM&#T~@w zSGqdnP|#&0Yh!o0jMxgQL^XM2H*fdOq0@MMjDp@+Z_9L#y?``4Q;BHW`VrH3MiWX_-n)q940GQaG)%A*WBV5no z=G8nYU~G%n27cx+sn)*3Q?^ZwAi9xzfU1u0x$K*MMFyRLOY9^B4w)v6RGDf6b{b3B zOWZ#c(zKku|82;8bLmF8e|7A-Dq7+i_`xzb2N*sVTm6Uzi44H~ePv3KqUp+`{}pTQ zHC(5XFtm`s{5(((j?SIOv_j;`7O&)20^A#^13c0r+u`zpi)d;lV{a3|QzdBI?TJn~ z@;t9RL|DmN!c}(>f3Wo9L)A`>*W3JkSv=b(!s`pgi-H?YCxM8!TLXq*i?IXpn)gm; zh$&;31*x_+kBj_wn^%a{&Q($2L=H3V!(6LI#xHYp;$Y1Kod??>X?~1+G2&VNyN7wT zIp0oK z9?5Rl#}0-TXkP$Gd4%j0W~@F$QZgrsaBu>Fg&SS_5zs@>;`pN17sO{88715Za4}>+ z#b5H`N$cn9eFZV@3((`Ia!=L+ zq%z($T~%^Gho(iNidcNOFC;;C4<61@d|>Yd`|*=~`_JZ!iE(nl?bzNBnFBo*h3>CF zBH^}G@!nst2k3gyH3w40Hsf8Lx*(@P5@yg zjs-{Xz5(x?X&RAAuFrfrg-Qj%8m~DDkAzoL^0NKV^7P!Nl(G=ej(Zwg-ma*yedDc= z{v}ZkoThc3i+l<~HND`NAPfn%Ug0rWx($s8IYPfU+<9Tuc;qWX!V{m%q>h)8mdI#XfOXJ{z4o2hP(O{dk|I#*vPw~0V%2YS(}^wuyMRT zvxGt@C9gtmiCiPu3BuI{DLwXe;#<=ji^Ywa4wZ2B9;cq^eRp0q((s4+5I>&4XZjPS z(377b||N+cz?sbuwq-iUSd6Pveyr*MpZ><{uXBZj0*S&zIh~qB_ViSX`S+9@hKt z4T#zi|6O2tR-8m1A_yZ3DGiy$d78*~(rVvD_?63}#W`N#Vo0{ji*%ZKdL!wx4|trL zuG-Bj6prNEemk3c4NX2TXV^ql;5ueq^tvU*HTUq$m#|;Yo{AQ@aKV?Tf{~0$Mt}-_ zUVwE#lzD)tL3|y$vn`NA{ykQfg?nfy_X;Vr*NWgV#Py8?NU*MIFsC2%qR^`s;zSoB zCPWCYvGbp4*IIaYsKNuu6SFC}H(Mn5d&XgL-O?DGt~ljmUghKbj3BZ?jw}Cl4n?M8 z&{9T)C^D6#J%%%ga4v&_lgD(!A$t91ps=iryZFKd4o7|dc|c_Sji1>?!qjeP#Zt6Gz9R=-X%TMy#>-a)wWZ_^)^Vxj z3vT|1nImXvk7e#t&ljS@Ei4EVSZ}N=*GH5pXR&B0+_JXz(2HIlYn+0|^W&sd zK&1|>(*H`F@*DAEa+~oF=G@L>6KCpM3^~~VLW>U%u!Vym)@xbC45<1P7GNoW1#1i{ zv+BFj`Rf7*A9+c=+cnqRtwy57frXA{cRDDHfGYq%Q7eZ`dpT6Et2%tTfkOplBvXld z?~t=djkWpsH}nX(;^fxYt5`qrA?sUkjd{nMu4a|zq3+#?`p3-%p6ifwF(3p|QfLZ^ zdtIWCt*yeWfSmiDzhF-H5*&k+53eðQH-Bf&sPsv>Wxjxc38eQ|70#e4$Tzoc;{ z1=|O|@k8iF-twdD@v0a@_#TR?(O)RPJsQNux&sIN+(!J-UBVh$>R5B9UeE{bbaDBK z;(;iJJPFzB=hIcQ=>DfZKRqVwy=Uz&7C!)g_KexA81i%~+6MO7JRT=}=|2tOe3K~M9ctW~|Map~0}tfw$FguR>i2CB{jB4XoB_k&F@b?iNX!M3~t4@dQgp_*h;A>nn;HZEK3!Uz4 z&TMqs%}v~T@9K$Ki%+z^RvXaXPq>}ae!U94GHFpKq1!SN*TBiC$EEpLehKOz$JGm{ z_*Mj_2X_%89<%uFbh_R9zP~CmP=xp9|FCx+Th3}%5Pl`1kcJ=$k#n9A6bM2@33z%J zZO=UR#Or!p-_U=zS^>`4XYZ=7ngZH92LWob03Eu+^ydT77RY49J3uzap0aC}>3ura z36HM>s@Jvj4Tdfz{uXCm{bM;CVhW0SxvrcId~`h9#!T5lmkcF{J7JxgEc`?$tvkb) zkphFEM2&E~M}^>zqceXD@8W&JEH_oj^H1^UQRDV=;TBHR4y*Zs7 zOdA-fa0556UyuPnuqg8F(M`KI(9|l+P*a~Ey`+7+K{{U-DRAkRI-k{%zOd0P{lkks z8?B-A^kE*$&sVd;VOz2meikQrE`X^XPp=~qGHxIC_#^!)XP|O65Mld{jJSPE4;;#f zLxiXfGQHvKrL|Jn4e4b|cFV6R1j$YN@acD%W5#NFMW4h{&9xkUL7YrZXwnM6*6tRE zxb6&s_i#YuI%D>(Rn{XjiY%w?<7n)(Rn}ULk1+JM2W?~ivw6@o@$3>Av~he)p*eg0 z#Ubd|o6twRVuYs-O#?2!3I!;-)8b{QxEtKVc_5${@XZ09bbC{Mv1LHZNK6@Ki&KG~ zF>8K6XkQfi2(8Xy8;~)8wF$JD0L-JFyu#iX>H=x_jI6F9eoyK>|kBN#qWn zv-nAHvLnb}EzT^3nrdtF+_$NQRI_I@l8AK{mS+o_nI3k0Pw&lQ!#1KbgOk3O7SE|K zq7=<{^qu&3WQ$(ZlHcL2II{)&MMq7}Y<|VeKeXCsRh)-SW<>mgOu0Pdip=5Rh94~P zd6Xs?_P~Y*C~TD@-c8&L*MTBB9uMsug*LSM8hv>O`+x@Gr0^3y=jRD6C90U zD(|LQ!EY>=jkZ&QM7Z|u^5bO)`cmoiToDZ$KHBVU6g$i6519?xf zp`?u>JyHi|$K?##uIP4GitQAkHG)Dw4Ob!;ZOX*o$7Ui(tBKREnChHWOBxlk!0(^8 zsuKARuI|8xf!1Pg`b^@+e_K>XRm2H8Vn)$(a^^AZ6Y?=(+Qhz1H1=H1D3CL`n4TQ+a}) zj?{D$3eV3DrBZ#x`=bv!#D-r#)4ri#P3$4_r*u_S>?EEj^;HI8 zCW>$ZZo4e%)s(`%OYtC|deFZSNnG@R&^P{LlA+m}Dd!exJYMg7DC# z4Rd)0>lF&d^Cg<%a4U=c{(dPqw}Ie4f?=F~>{iRk$1Bp&6(ir!mS3y%wty<+Xk!a7 zQ)s9y_k9LN2MTh%P>4Dm7US}mk@`2^$>^jJ=#c7s&xn4^j~w3BH$erOYXKjr*nE{?r~k%X`jC0UZgwl1fZ;q1o&xxm_-By2uneFK<7;yV1+Cp;Rg zd#3a}h$k7*FG9OVqK4hOE(GcBuH#D>?o~WpD~lQgq^FHfXK3l+F_PMH_c61C694XT z{APm;4#a!ZmHjj%NE9@@O((9rPqquK4Y~TWu#6?MUzP1jlOn0P5c0+NH1}S47?0W^ zAUzN(u_cQpT#_xFOuYQ0UE#;8FT6D?JqiD#jNm-{LNM2q($>#khthrG z@h7*$S$O%LUEKS091g2HA}W=em&JE}zriozj0VYDvKO}ov6$eTrhwxQ*%&$bOqJ&1 zFoymze*k?7e-YJZOfreGL=yT34l1FZH$Mn zg+^%fVf6a?9;)k2OA!aM*JJq0qXt`js%JRF=ggs0*J%L8I|NBNQEq&-7SH5abaCr% z`)tMF&T1w{lrCLP#qtXm_ebv~b%xa5$fR^`S>o+I!8!lqCPi5ylUU4UPAi2S@37Au z)h`Pe8@vgkE!Z*ide;(E`|7mGXgfw9VBfWO}N#A`XyBCP(q5?+fYyVf5&52q2ci95vc>S$de}1jYSOsmDEAM95~3d$WLFt^!hvdm}Ex zE%Ku?LjTyG{_a~!wPF_`XX8~%Tjh^2?tg4>H;#n6)JEtO(YAl^h=BsIkH^%29H!G7 z#p6g-7fhk3wIL;>r@J=VRvF15tt@AUTdF`N1NAtRG(|rVLQAk%XJ7;o$`ve~-`eVLflte81#xJYL&yV_Barw>lo7hy~Uk`Q%9E zeBY=?MquEcDa|hPOy`4p_1r@}IHkTfVo{x-XRgmI^t5yUX}pH8BPIr@U&#(LH?Zhl z`=H9JM~9x~Fk60=Xy4VfQ}_sWGI~M`lk4Ac?+%}rrT&5cd%3 z7S>W&ONy_59+fZ~)Ba!sIg6=oo5^1)7w0LK-b{>@!#c6+u5=u-c=96gs_4uFvFrpWNa0GCiV;3$qw;n-KG+aVD7~ z*>1&sv3#j=Z>P=A+@H6klJ0N%_FlTz?*T}xXdhr=bo`mNF)LGD`?ub0-~}G(!|?e6 zL-ErDYo+;UM1{8%-g4WuCpWw<<7!Xin!oqCv6&jve!slt9Mn_~`R57D{qA(~Qtvrc ztLrk?4o&_Hfq)S(wRN675k>I|lwXS1@*Hty#6JT$;7+CK1^#C2XiyAyaO&h4BU8^I z?6zd-NSkk?)|%Ysiy4Ce=N^BFY$7vEVt#1(^)|mI?i6{K+iaC48s_saa-ao9LgVar zx2?Dt9sh$*qfimu+yO~p(VH*K6!bJB>19r^jf#dr^Z2f093K|{sD&i)J0U&FJ*orb z8|0<r*`-a#fp$&c(7iih}}kPvM!k2 zgTp+A;D$8UYi9(X5);F|szG@Yb7=Y_l81GT?7RVBAXG-zz{Pwyq|$)3%LD<3fE@2+NTMe&@Pzv!ONJvz`%67gJ&DL=6qi?*&+anDKj?J*{yh%` zZNlh6>cQtrKY*@xc)Wzj0R=@ueh~|c^^3PZ=qrTyu_Z0m=L?+QoUopi;daO4OA&2^ zA0QoX3RC59y>mkEC}VJT%5s3nuc%ot!rpIaCq!g^&W?SRph{OR^SY(cZT|#PBAh)% zLJ-a>e-LDpQXkm+$7HzmFl5c*+RRD#_n}bgdWE&bgREj9I>LJ)7(IZK|27u8dJD|m zI78s>6ys6tW$-~DB}|sLZFPlPb+5n3BYC}@U~XLSUOI4^P8zgdQ&(_LBq-vb+ZI71 zLY-*2@ao!Ban88S)v@l}Fxyk_$TMi66hJHtZu3W%)d?>QuXppX*<@wkX@(+K2=Atv zOxm%<*2fN0huk>yW?*ne=#erA2m)NLG>kW&$`WPrd#m3}k#?^ayHie$ux(uH#t>mD zAr{wv;cU1+ar|HGO&pPhW5;~fZD*g0H}b@QUpfrwG+Gm=q6 zR@WP;aD?gb7`(x~U7lz3AdK_rRK4J2t$}9qNl7{{EIdO8d9EF_VcurT5ARbnIav}K zCIC_1f?B>4cR#aO%77ni0jVV*IbiwB)J~wt@7C#`o*F$2$AZrW-GF#W3=LsMdxhWa zHB1ZdzH(cG0sbhu$9TUnZuijEBL9$l`njJ^`_>DL>NB~l!1W??QG6bb!7I$t6F~&H zoglVz%_~$$e9%NSiLjVhu4hJg7Kf#@nk>)qXh+QQarKk4hwM+})s%Ys&<1v12ypZP*meH6b= zTPJIA2k(%rHhDijz_|;gOCi_yGx!og)7Tu2FLOdWEgTsMq8*=vH=?|O)FJpVFO{e1 zp3K6Z%lte-L>YO+YFbI@p|D50r%n{}q$(ii&JQQ1%>dCbZ;#qS>YhLS;kvrtyd0Sg zpkIUR@1Ox}4^I?36|7fC&3t7Zn>RX6a&j=K%WylT(W#6<##*x8k!+tLcyg-42MGi`U^bCZ>kTEF_q^;O%tkb zCI|6ctbaRU)T;#0TVc4~n1>NKaSXk#3^aLB>VNdN5OaL5A=C-=$|WK7tFv!={LDGo zQ+$W>kS2-#`Q4o0mc&0-zlc|5%5FX~+gkXvq8qReIOPX0JyJQQnDbK|*!d?xBZDgw z=}79$l8v;->{kG>4BoZSxO`!N352_+uH&1orpN(gL;F#2Q?aZ09NL%(A0X^3;TS!F z`q-Sd)*6Xds~`jF5Xi(tv}>$BfuO4?cRzUxXAT8$My|vp;D<|Io7R~Xy z+{TS}&yZl^?Zx+hc3MgC5;I4ukoC57h9dK8n*z!|mV9-nj%be(^k=IU{vsV*y3xtR&j!4GGv}o(jt2q33pK%0~8` zPwFA?Z-H@a){~T)@sasV&jZ}Xu+3}gV|$;`3%@{3nZ zjQ)FU)M2Pii6!1elt!Dq$xa>YR@SQsO%V zM6H<*T^^X5CxN)3!rrpcdLz$j2wvKC=O~Im9Nk%YP!}w9^q?YRyLbm9xI=l5 zcp{x;g-)5SYeuP4`6QqtqE#axej&-}Nnm=1KvY=&06Dap@F)o>{h{=&G2N3^?UGPh z9p`5A2r_6yM=o7~7(lUV5Mu(BSm%tz>)MK{o9gV;z{<#&K-xn{4~1v9g)Y(mKr1)| z+NEfbyY*@F4a~4KZoy&~#L;yAVOKGsruwOSt5*szLORbpB z_Va9A0=KMuI=oMpXl62a2b?#W5nazlCMPq^Pz!L0e-BwU7^#Q~6s<-9%InMV3$BB8 z8OJxVuq%#|UzGI+wL%2bz*|ytV?{s_;TgWN4x%Ru6MeayD-P@43bN9o6s+PLs+_v2zP={)mkCua2~xLvSXs zK59K=Sj7B$*R^K0Kq!(!gEtld5XMS> zh{IZFUMg==af-C~6LY363%HqMc+WD-+6*@Ljqem%g4{PZsDAD-_Kn9BrK`RVu!2b@ zIMmHINrNO%QLpXvJ|CfM(fTue=EU{l4dw%$@zOqmgef6c#lY>rD4~krHLwL#WKR7t z4*Zy$GUxn+c~`I=_HXz`SkuU5xI{5^&5`9?Pn7`o!fTrHj6EZoVTLTPVqv#X&FlE% zrIk3iR-#N8zQBLeoT&x9@0i{DL}_=#;Ky0Q9Qfeh8C@^GYQq5gkMfa!GiNlCU!0F8 zJ9fz4@}M6gXcoyJkWAmyldHM97Mo}v53lr2w0?gHF^Z(oy1dQrl28SDJ(L|yxT14}+6bh~q__;_9&E8+ zJsW53Kdu6upjTc}4G+@!D&?^gaRcH#Y*gfm4?Y4*h8mFp+&Cjfc&xn(UxS*OZqfZj z{d9aqH8wzdqxO`3OrWh8^ zT{i%gP$kq(b2B+33%DiuoQV*_C}`#0#%UXTqF>7>#qFnMl-=Y!8oqeq$^lYx_vg~b68i+5uU5|d#rAE1a8B$ZeLIE46TXxJoQ?Xyh+96lx$^+ zvt%D{I0T0jnu1VZLvO4=T~q($&OD78Ypv4>$UqaUTU6%QhT;xADAdIV~0cgCr1%NR*^NJj#5gO5xn*m}TDNsw;^#DwO z3>fC7e-7t9`l1Q}l^S`zY%}@_)W0TSW;GldI%&IM5Hh%lyW(@;5@GZ5=qeVPJzkgT zGVwhQ8z$=Y0-PD&&(LaET(d}5ZwzoBVR1Ys%U58)8{oO#z(zaM6#@mhbR)j!TDwgj zV@j`NNI70y!71*vB-QDD?#(vAjv@CCgMCr7ws)>HDUK zT9hP$t&lTprRtJ&BmkeH@+K5z-u7h;f9-0&|0vb|r_w%bv;RNZJ~ZU;YW~so{gaNt z|D)~uN87g#BL6!bW1IKpN;!p2H}fpJTXEsm!=-7 zzI^@fS$<2FARMYwN*4S~PYMa9a`+owL}n5yb$R#f^MEkBm_E}*b;)HrsD1rx?iCDD z7x8eT|`OZIfS4M@eD`-W)+89p3l!I2q07iH1s~NM3&d+e%DZWgO8P|PCt|^OKb45u0o?|@Ek&rl^kypE!2Lnt zW}+#8KR%H^ zcS5T4@v1{_RFPg}ZtGY*N?K*GK9{hDnfS+!1@vehb*JN5zN6>a&&JibktLmvZo%>EmKj4!s7 zDjMc~v|Y^nN|jKq1)V*QO3MuuZkGKxs|=$PiH9O9k&KzLU;b{bqElNqBRrS@IugX% zsDL*6@$EjN!{GXUBgz8qRoI}~9@Di$1Y$p{p4UZd({E6LyP7Z~dFJ-{j(;<^cX(|g zij`n0E*|)r@aUrbZY*8cse`&_E9VKPN+Kg6o*qide7lkwjTy8e*?4IQmLv5RN3A)p zGFoJl6j9qQrB{wCi^Uyws{kq4*F^O&qib%y#aliCHZ%8-_80)^e2WZ9r=kwQm%50w;REau1i#e>V{p4Y?hM)Mt?@i!a}B%)Mk@=Ftc`z0hN^ zJkE^Rh_m>SpW>rBK>IfvhONauLQH$;{DWas8a;#w5(o06+ZH3c+G^&Ll#c= zIvJ#-tHA%$;`{xv*VUe0vhniwr>6<~D*^9eIAlVH*uav`agqw7kL&F#f7Tv|rO*Ll z@K%;9_v_)Q>{rj=uh8M#*i^-T)PmIu2lkXc^jd@*6WW-l#%Mb`!TQXuc}!oVLvVsc zudG5vlvY(2oF`xOM;__a=om?R|2>`o&lmXpwnejq@s8ldYl#Ns1yVP;`C2}e_4u%M zl?v$ey|&L!yW`b<*;t}sMAGz!|NiPmN=Dq}=nk!Kzi?yRN_~IA#cyAB@r|z=VAdIk z$y7Z{`PnoIaN%F%fB^e!ar)aQB_DL2xekWMzuH#MXY=oP4J?atl?cc%Nu<6aZsUr^ zp|mW_HhnKejl6gCk~Z0{WsFGf#&*X_BqDvIixzRZ9>uWkpKe6tcA3iUjh6U*(^~df z@N3|wla+EKmT+;g9;%XDZb#ypwMs2Z^zr`W&dqq%y&nR5@GYa~D{=-IK7h8!M|hY( zVUXMpvKs9oXg7*ga?#_i!I8#;Unj7Zk=vTRIqz}v8^2}_W5;7vz}3xN7%bJ<?=FxUqz?yFE?Ihloz(%_(ZG@x|RFK%!G?f1k4q+~`~WjPFGtF_`#^M?M+~ z?shh7BH>nf!kGha@VsCvQNyROZwS1)H%c#`20B@cP(XW?IiPA)SG67k`2qjgq9CYk?s# zY@q&cf!^z8ZamwzYsLHoG%9;y#FqjwSH`A@G+`~-8kXvaUikI%wT}Et%IXl_V*7{b zq_U7w7vZ)Q&cmA-2g$$AAM!}4(xbHm7R#P6C0b10St5nb4u^0n>*nBTKNU89Cd=tg zh5$&ymZ)M{zjo(2LPm;jb%FMSrV~$jzj8eExloHZz!LIk_Rrbvhe{WES;Y(T@8ZA5UL?_Q zyi)o!mZQ%8Jd*9tOg+D(j;4YSnQ^jPFsq)u_ia-h+{4vq?b}+l6!LRdK9{C2vJ~Z7 zzpCjX%d{HbMERzp(Ku2Jh>=&Kd;!OfA3os zYWGqvPVfF5;zPV&R_4gw^ih{SkDrlG@!>p?Ju1W>;&|uyb6T|Te$__M#zCd_+G7Tm zy{=<3WvRyWc$`lO>;`BRkFI6B-p8dmRPq6d_pZqSEAY2o;v*D6z@^4w>%J6vI_?Gy zo4GHgevX!w$_PlpzTO@E5%0f^L2HGElKrVg*SZ???b)Es-lF6c9_w(cp&EI;IGn@i-fpmXPp}i6jT%E;#nFRh9V3}n$^?e&DilVp5B+G!VF_;=aX8G(X zKa3$v%=p~=J*&j3;8MFa@0so3otA$KTAPEIuhjebrhY{Ws`VOzCaZ*2=ZnVBgPbuV9qc+yr<)f78r1b<@hmK_?emx52kQA zJ`rwkA+0|3>fk2Z)JE3>ZLAn?ON#Zm|D^uG`qlbEg|pUY`)4Bo?Z{aD}cl4*y8s$%WxJvL~i!Za@5H$xi6*E%f}kFRjJ0k9~r)oE2qf76)Zf?2Wz#2XRWr{ zHq+$c{N-cog#4eNKj9mXIw%j(kDQeFCE;d>Isr?~R_O8R|0xDcpl7 zKV4oaMJ zrpGsr1EYitwElDtvYC%dTsjF3lZU+`lQf&!yXu4^WqhC|r)?v927uCAdOt(}r0@H) zpH=HGI5LcXEDnbE&X?cfWzbf9_9AE^2P+F>CI~NUv6i~>4Q&AFXY1GPhr$6tCPb$3 zXa2omm%i5ZqYKk)4vFEauh0Z7Q@-z6$zRF()luEX>oLz%IByW-NZuRh+kM@y>op7R zBC*W8Y{7?z3>17@sRg2a*jO03JH4 z&X*rOg7V?J^1KN@176Z$q!CNOv?-KRfq*!p-hC zzo;~tjd=?qQsfY`Ri|xA4{t{J$$E#rWYvj?!F{sqz{oTyUy@f*Nu((8P-v+wnHgUN6V)2;-b!T`>-0dN zNKEW^FsiL7*ZUnn{^TzcG#ZCGs7p2E)*s$DW(D2E1)Phl&P!wV=w{5ZdvlP|_n@uv zjYp`7o+LV%+-qorvDE78qBFl6QZ?;6-66D%D-uW((-Z`kzR&x-ZAs8i%x^MCku@0% z5hu3qJYnYqGX)qT7Xv@B>Dju*QFtUORM>1Js2^RP9YUHSVgu3cWKO{aACs$|%@-p+ z^dly%_F>FQ`)5Y+96-M(MW}}b8HeWw_1OP)JXoK3tOEIDy(dB+LL#jf(_-%52hQ%_ zkIo%Hbx-JeV(2=GSzl~`5twBDga7eIG&UeOw<2xb45Ch|M$43nq;ohFQL1*w^gt?b zXbO;1f;60|6jSF_)h$u-xW0vbL%AAXt;SBBJDH{7Lq-mmbV`}auL6!@r9oHg}OydpDxeKJJJI6cOD~ zA5;C*@$OGxr2T4{iSQaM@j|O~gD4XIDY^_yTR_jW^!a;X9`}&^vh__EabuwP6y7LZ z`O$e$Qpag;ceRt#j@pBxVqD5xPl{d*O^BC^;h( zkqp)4`{5j>-M)Hyy<(>W2`QW7_>+h(m4u*;03nU%aKa8CzR`Q0g~TO(sK>uR1OA;#h#!2$Ddv+|A_q&Jbj7sU?lw1%9KAfyHk9E9sV^C~4js#o~mgu=8E+H?IDo zT{_C11&`GR-dOfhY73Kt2E=6E3!jBNm@seIZ zkK!aG=B&C4viMWu<8!sYz`(PRh?mID5z&UkmX|rKp@vktT9OFF>Qfx{8 z^+mHR_khy2GWz1}?vJP#53xIJ(s7=?rtmEXN)zvxNI z&I)_Gm_&NA^1X&UkNs=7Kn7!<*E510?L?^WKcVkW6JL9L_tl>dw4S{uhhK;GP(xyk zZk~4YJhL0Dq040DL5(|Kh#&G5EvrPTX`#EBDCiv0dhmljl7)q+qcqe-=z!ff!%|Nc z>kzqzGopopM#ARfq zkmsTnCc+}-Z2OLTliNWOSIh@mtDQkk6QIx`Mk=I${*E`L_XQ!yph{2l6N4ok`@^qb^TWP1t4 zIP-yPE9@-*D3Or&o`nz0dJvtQRc?#c-_eXRjoLScgBV?2&-Lc~^vfu?_S-S*l;X0z zIg&l6Cj(hm6^$xfnWgo-CLnv&C|Dp=4H>=sxm<{t0@ihESajj_K`>5xoE^kOKAq?N zcL-8Bk&?kc7^RPs4fkpS00#PV%tji2Rf+g!9)Kc3<%*l-DizbpPwBMbMiG-t{5 z{P=wJZ|k>PMX^he3A*`lGe!B&b=oHe|2O>s;rf|+OVkOY24vkx6SkTCj2#Wi1buA6 zbh_F_k~L6J*o^l_o{83~K`+R!msv(1Uuez_)R3#Yf~uq(*GsDN)$@KbtFzYM*kg+f zMC5CE!NgK`?udv}P32$rGg?!D#y;dKbetnCV^m1!bb`(tRcgp2D%3Z|NYSb_9a2uAJ~5LQ|8 zXn$|FG78JI`{c#Ce9^*YY%YH_uxercfB^H}S@jJWl<0Dw_R8XSHugFKouhqC{|Vk) zZr-bOsqN8bHM>C4d#|xbLBXWgQ7b3Gd7A3%V>StrW^#XJ(V&soS|9I;c-uGQ;9^0y zGl@ye3E@SkZY+DlUeUZsgkL@9pqd(_Jm2y9&zytuRX!9}`q{)mD5TqGnXmrKhUoLT zFM5;L#Uq_v7u&dJpgF=#@mHX&D)xRcjbr)!4hUEvu!#eDm}JDnLZ!dx$4Xp(Q7Zfd z8b@OxyQ?YLP)$fL)7k@p>8{njs4?Xo0lR263JD$0?wz#^pVjj2)iAPYBUt?Y#h=}* z*XRAs!lrE|j1eEYCGd=d{#xHY9_NQI50qnBf#$Dzom0PwfU;3C_nD`N?LYY}D6^RG zTZ3ksE%bRCRC#z_kNV}eS5Y5}xqShTt+Ce;AK9cOuSMhY_B`m#Fzl^^B)Z1v>y`5@ zdX1O}8g~F4cG4M9$ur4iZK8 zT6dbM>G(5g(Lp@S_H6STD^1wzYI`%P_x%>Hu9WA;_FL}nuIyYyRZ+Y-siSov4wu$& z-fUTsDpS>UrN^c-j!tcI7QksIl>AI%s#KVSGQ> zryC#Vd=U}UuAozgdpN|H9e%dZmmXYtNCG>*K3gEf0_sAbL9R*h_npO_tnUx>V)_X* z0iU>FKs)~Y>X^PzrwD-(Ob#9vf4ZVfMGy1#o2>h2fya}JalB+$~M{Ppa z0gZTJb^=5BENoaf18vJHC4!x*+13e)K*BWcrYz1&uqA=KU67PjBLIfiD@%|k&~&5VdG2Oe1EIO@Nwunco`F& zx~s9jezaso+;E4&Lo%Gk?}6jOw(>#__r#zb&F)NAs&c(#ErXTw2mItKzFckI`1wz! zmjE5u#FO)7vfh@7?>1A$*h%^6O3B7wnB+^$!rOKNsWNt;Td# zp}OnU$GXW`#|t0qw_i`yYU=eOaT;#p%hSHM%lv48^Kh@}O8tuS;={RiY0y&9i`RNv zmGT@0XZQ2VCDhU7X0MT^$H^^?fH>d#WRW%Op1rOLMLFl-Kt^*EH?L}f31EC~iU4;{ zA`Y#3MhPY)3|Z2b)FBc&tU!t{FPAMS`gjW8*l zg3wDX?Tsp{i#`jnp=vQjmUo#3b$o^$^aFW|yk!8qqr7E@2 z5`MO%7Y*W&dt?2};WCnCQ?HIl!tOvJrTQ{NodMZM;=B@JeJlWiH1?cy62K(v!`@uT zVl!Qce&_+>YKpP3;bk+&CXxPkmhhr6=~)kd}~1adA~j zqrTyZ`)p1$w@~EzY_@L4RWjnezCSNh;%13etk#rFhsKOJ$wDss{*jM&gM=K3?YxUc zPo-|LPXI=ZX*kh2eQns-5U=rBj8>E)7ZiM_Z~z%aAu_yNYRpyr5XnUtyq@Z#5f(UY z5uoZAaN}SQD!4>q8{^$&g{49@ppI5TOPG+=V?t)@Gr+gtFWzf*%(MufikRXR=r~-E zXcqu5J|IJxUs{~8T(EZ5nl%nhwOkK_vM;L0uRhLRN;AQV!O?cR8|9lfmMquz1CB-w zol;nZ!i-%ZTBs98V}o|7?Jv-`e3ZCEa_n+%Fb)qD%+B@=ZQl1T+`@@6*gafgF7xi_ zqUGmbTXB68n%3G$2#*GVnxsV;#GaH$2U1}K%*&O8Qh7;-2K;!Y`g;<^z)pZFg$kt=o68B>e?I(Y2Lp{L^b+*^P zXo{V;>+Mp(Cv5IlDP~IfBRg~O`dvxA-+Dd>Z8Sc@JJon^Yh7ScF)Eop-BbDJ0=?lq zY|cJ#kme!V4}()){jlr-IobUon(xefJ2pgg?`}sR|L5zhVZ>dzhQhPjlB_>7f9xMN z;;J7uZA@8jDGqNrH%4L-ET7ZM$qv;3*0k#iVx_p&h814|aG_1wxt8TBlF_cMDlsJ9 z$3yl|o^vUp!7>BXV% z_BZ8Ck641Y(pYCYQJ%%|gv2-xDDnx%7~wY~@8W!1R^-^osCe;8qZi2mPB3*BZ9j46 zaU4GwGIbLzbG{KG^wbq!HlBPU22&OlAJ+6hO8E=a>msw&f!*Jh>BNGrdnc_T(VF2k zm`;1$lHg(2(1uiXS}aGB0Lb@x2EE{LkErPl0UqpNq8+j!O=9$Xwv}sH_@LIYo_2HP2=%aONou~&MS(O6TX;B+I@5<~eDOxL#`WOVkecmi^()i&U?4$8R_wju#q$I7fhwE%)QT{ z5Sojfzsd`ZB&ex3LkO-4dfKcEJx7>KeDgP-GRX!2=LE@&a;+p z-FUM*_HgTo!PdW<{N-iTn8-aRq>}Rr`|F(_R{26Tnh?JZhcwAzVvIEJ%Y4$O;E1P( zg}I}nw=VqF`h+lIJj2!*&BYrBb2}=BR09R)eFvqDZKjXXZWyh}`{i}PnWg6F0G^mH z{oOBtlXAD)^~YLguH;22hBh`MGY@H%SO^lDbc)qz=%+1z^ST8SiMM{p%)z16-`g=* zd(w-=TFtAi_cOkqD64Le?BKnFNmx3s8K-l{q~EO2J`}#edAmkqnfKRl16KSw8G&vd z(3_}|YS|y}vcXBMQiTDnq^6!G?z784)7J+wUF!&%S3ndC@S**tAWvPc4@j^8_h`J3 zdx^00$;KwV)7>iNmxq%Nhe`;g={J5uG+9>3@?`g^AgcT8zHz?Lp{L$$*qjn056SU3 zygdjRGn(y@z25mJ;6D^GHZS6hXjTMz29LL_NVh<;?&D!2lL2GE*xJ9*@d8mD3&V(vSZnDzcx)0z0cCf#Jyxqs?Lclj=vUS^H6uSVo^ zp{>0VPwOqp{a`~qN$1vhXl7RKvXRImF&c}nhKzj0m=aQZt-X&NZlES$Qh&i`Xh`sduYH zC%+u@d8z=9y|4P%dQHF_#)strcDlMSTG>rC-SaL4Cu@c?|&+KHHJJD*$E{jwE zMs=ky1l*v6bO&gPt}(_+D)SAwQuUV42|ujx-EoYgj}0$t@rsPifAIP65;B@;;ua() zG*ys@QUwH{uH@U?;iD+ug8+~_^8Lv zH*4NsC448P4Si7dgP}S<(Z4bqYBYGb>G!d3(2%6y(5?G# zwoDBz2y)`{(gRrB>Pb}ILurB9IN#KR-C%)fFtSGd@jxqh9|39KtA0ZG&Zu{#g8Tis z>pf2MpQk|Wqi4HFYEifDbiz$gZ8XB806$>?aOUv9Rzl`|xJh(sEt2GU>Q zcYVF0bVu(C?Y?5LAJ#t!s)s=713)$%{fdQHtG0wPm!!otx#}%69#M2q zACEbSkaU^h9zLUC(SQB~XA*ulgm_mSAr@kP?-$JWQ+sn{&)_R(#wIEJi@h^zQdZlt z_Ajak=>#uD=@pP3?&wrnX%+bOHOO=7i}<20>aHRxVn^(A_Pg^<%~WBnHRl}T8OKH5 zU;ZR0`WbQSpg{M@G&0VyYK4PUbF^9^QbTrg=n&c>X8O6_q=SEKY}dv_NCQR=soaK} zAzQd!=cDnNPDrK3kK%BA&Yn?H>6-5NM$SjBapmtkO1yhTxEzCrwjiN>i{PQKDvLD{se3Ap zqr+cC$w2XzLuBmI++lkAZz=fR_k)gaUiLjTkKYoevlBXDI++peRr_GKmv@`Cb7&~^ z2Kd<=)A@HvQjj2+^Y(lCUl*69DEJMf@5?-0QR0CNVJ4`s`Qi%!O;kbihQ!yN6XRD- zRC428bi`+##RJ5xpY4O=+b1~YrKv1gfP}1lQS^=D#HaI+)bWEwZv-`xU$t}BeFe?2 z_>ZIP_sK&xWyE=-uXZc8dH=k9rXVZ(A%j$|IKO+izrahUe6KLw)SWr%z=7Q_WCrS5 zGEMRmT{#o{-)1?Wl8v|QSP(?|@JDw+4h$a4oIj>D%B_*V`gczjAXI+l6Un?`Z<>;8 zz8^afnFS?JykFUMyj*F_?}p2R3fkWJi)7@MviJ5G>(`;cPQ+F#bG3y5;XMJJ#aM6Zm0?V}IK zn?J@t*wo=5=KgIyB_6u#=d5-pV1W zlTGs0OUhw4p3QZ*jk-m|-m{*w@dl;v@!U9apD(nCsS&#-RuZ$=(yO_{d@Y%@`$Vdz z5{J%xExI@56rLYO%W~_T4!{kSvM)nCC*jliFwUol2qaCNczu&uJIyCl@gldGTZsd4 z^NRKK$Xsu?#QAmN`Rv>N)qQ7)~(^Kag|k@4^Hx>$Nc;CzchYR|1h$> zdRc|?)e*Fb{^70l07xlE%t3_HAt_zq_zr0S|2~0R{)w?L7z3Hy2;sOl_THD_S%}KJ zuh_?U6?~dM4v)5| z4VV3nw^s<+guKW(-RuuI7>5h5UMr4St)RtSj+DTUL)}hwj|p2k=u0JIPumx#C|HlE zY&(P|LQML5+xFKuJ3$G|hH9qPZyDyQ3YFIh8ZQVFguFaL^wvQ85#9xQIUQ2sKCN_u zv{3?2=d(5P>V#LlL5O{hbek)dwG;Xa$Y`0M||=?@hEfNi1|LJl~YV+>(2}$orYjrPW^-o)EVd zyt4Rrx)b+!$?YW|=Biwm(ML4ds03$4$3B(#74fRD-2~ai4cd>%HIw%oHOJJ(v{+37 zNUnygKS6)8pAQ{#Vt)0(m(YgPRA@EqB@*dxccOh*8+G0>2cTN845WsFApzmGA-woJ zcn~G&;%@*Ue70O~g7|=k9vmaV@AqzCW#?<>IeRg=ydh?r@)!i^Z#G@8@hWAHma2`j zB%2`^*6iLUzn}0)Eo`VfvQO|aa&s9_fJ(2Iwpbs2{~hhM7g4|Rd&!Fah(ELgddo@H zh0@fJcgmH~_AjQeUk`U?X2<2%s;!ff_p!sBwSPW}G=C3)N3$g()XE&Tj0 z-l1@LBrlxxG4hC=I`i*zzs=|as2yHtLpAjSOz>|VeH&N6cV%*TX$%*QJYtgu1p+{1 z-69E`Ew^Scs-tlNl=$KVm*qS1QAVF73$fU}(LVOWi#l9Z{RiCz+Aj!a?7Z!f}a-$k;ze*Xi101pA(UfK7+yBHL6yhtJJe@t1Mu`6gusM}j%YukoCV zrvC-Wxc-{~!@GEM*5%QwbmXpx#;uTB%spvJdY?BG_)C>5s9CCUc$MevMNoJp$Q`ZV z4^gQSDxI;pIMf<54xJveTQ3*(c{mXgqaLrXU2{;!&qe>rC6UWE%{YvL=JR)w>nIy$SFF4-~RTAjf$*Bqk4dB%DlhG!TsuH zCSNPo$0M|Po+bgK>j*J@J84za%*thc=CZH2yVr_8VQCi;HiAF?-G_QU`kRZJeZ<#y zqtQNo^bq)_Q^O#^SMg@5nuDpo{YhPEB zc3L+0isxWC!p`*~xa|=iId{j6jY}<>GJ^yj05tz3n;+bT@K*5r35xRPHe`vi<%nD; zEt8p_&GP;tr4tbjd*dmzxA{|T`R7w6rHK^|KMgp#2g{Y;64Fuj!>p$5?OcPin-C)B6l(C8$s!2 zChd%MOmFlrdp|oNeX7ETJ~?~a#vgf!T$QIiIg7i%!Ja1j)*;+e_`>Wo&}ZJ`a3L4^vB^QR`!(>^8XeQ;_Z`X9(benkNte>|9lME zJ9$`SKH6({RYMyg3>JgQZ;AVa*JDoHuKKw;>^#!@`|HnpsjHElyq=8l>>1gCRj%7{ ze?q@{k2D-ueQ+A7Eq)1F;eZsjiiWZOeyvyjZl8OD5`oulHbG8|HmXr!?E5!|FV>k% zzd7V+U%5Mo&7_{01}M+1lx?^5OJ7UwNSQAvU~Wm=J!ykx6^u-Hq9cXmSz5RFX%Cxy z@J%*KE%CF56!KAVxtIh!U@72_2a+52J6VGa4fH5w8|xspwNufm^RrZ}-%1jrbUwDc z1dnh9A1sx29WsQkToVR~9CG_9NIATsed7@(s4tv{E)#HFE;mz?^WYldFYu39;~Q`L zcj=J1;9oWSj9Jsn`qi<&TNTpB)PtYcCg6HCe`Rvns%xWB!o7$fusbt2E>T?z3GgX7 zT<~}dB-L9Da%zM;uBZ-V>0-5KP%Q>AoertR>8$n9<8Sf1UJLY_^Hl=I=rUF7mIDI{RRgvl6uM&$in&-s70Hzj|9i>O=o^H=H7D|6|DJE|(cD25GNjy2;@BO|une$qv8E-&`>c)@)V?yH3{ zP>XwL*=tWsxuuCH=(v%V?sx20s6E4y>)%9d*eL0FEpKnq`-beUw`4RXzJc(_?;;(5 zvgHiJ#pHg!o>N?1I6;vNq6Sz@V&$81{30M6+9IDqRe@ohn;ry|MG{ikT;90Cn4lf- zNHN*mAtq&j>0B8`*!<|jT-okE!|JSc0OKtux@uM3%NNbtkkfz%VOi?#Qwc}^vl{!s zARpQ9^bHDFnvk?y<`>Mho8ARsW9=n$LvL5=d?fBQxmZ4VjeqL}mJpk{x`~Vz-(P%` zY0sPDrG;ZK*0LJf`T_>z2c&k3QCUzE`@sf9O9& z&>0bbQ}=9+YBENf@U9^2bJbE6dTHsktv7hL;{A(=?wA&9P$*Q$%Xt*-q7-G5WtlD?8HU z|3@n$IeeD*(&1NR)BE39r_{f{_Rfpf`@aHE|LcqK_^=|8lxe`sZU z2kyVq%HEiMUXM9woofF+(`A-*``X+kP{Y^)>LT%b_O9xF=0k1{PXma+F$gu=YDA7D zEz}9>QThzhL8qR^>#f{z-sYqGX-3$%z-m-;RaDxXJoa%4{6k&6>##6f#6un1G5*$1 zg+(}>(+!Q!^eqxhm`FiyL*$}j-=+g|&o%VEx7gq?|Bi&@s^w@Cu8)p8k#e{$rSO8P4G;a&p02YBB*}9+VAxaZiinz@#>|IS zCTF#ll@wOgBXoKD5HN*V=ITQ#*FU-|eRADr$WCIAJ~qK-XeKS1it1%WNI`MuQNmy5EUZyv~4Z4Sdng+G3~MR7*! zd{r$>NfpBzkL5)zh3i0Ex|Y;@JRllR4W%TLPd{SfNx2Nc)9oL7C^5Z ziIw_(d_;{+69Na|)LM$!0y$O9tUSqn;HKj($^&f=qKE4i{cbT7(z1^%N*Q24A!U)g z=}oSxCS5L;ek<;Z!Svd;-a<(0z~9xi`w7{ph%BRYL;a1?h0-7RJ<{~qjoK|avukwaES^i9?Lj^u4TU(gQWiVu=WZcxV@ z569?L_wc27)*2wy5x#DJvP9pp4U!|xp9VUYj#n|AjNWaQ@J@g+cFCy2cy7YLpb?f@_hi{3#2S0>3 zj2@uh7HG);GkJ}9g*J7qk`a z{NJdrcBL8bw#eD!8NE%AFxCl@fHq`385Ue26;J0^d;iFWBbWW&K-4ur5GZjDPgS4$ zUkEA)n&c)=!HkI5AddS{Q8_VkW*t=N>MdXae_Z*7$rt%Ud&WNjL_^>jsxT}5kxG25 zCen~xhDT?&?CM(c?(e4x-$zf^D8AttVU)pe>_>(0m@R&0!_O~qNEH*z^`7lAY z{@y3XKOe_8GGhNyoN3F$pa)NCPX1KuzCOWnX{T2ZEVV=+ksNb`Uf_7BKf|QqP!t(mJl~CiFoxfe5zPPSNwK6<%8XFyPC+Bb47N6T-H)xNy z=-@b|sjG`aTy3%==FxXO8nZv-@X-BUbdrX%=l_X!uyEoJX~op5ypIfsKvpyD?aITD zBbURI`N)@W1Mvv{Qr;B83(e}$&g$jM>@VhbncTX@xJ)`N_rHyuZK|1{ik)^OnufLS z00DjkF%(JCeSVY`H`pv688UKHiks?8(u^Seo(_aJ53|9(F#`&JXu4;-Q)A!YPc2eV z@GLF$65xQLrP)Q>woT)Y%>X~pSn={B><8%`L%@nG5ywdS z;NV3DL??&{flT1Rc~+aT;rkwLkklZJD-S6j*5PnHB_Pq?&kv1RkWYx9E!~IuuBFDl zE#Yl7Fi`-=K@vZdW@kg?Jj@IEFtjh(o2lZoF-lh;%~&scjnwz@MUQ`s7MH3u$-n5S zc3p8<@n9bH2);7FsDdMIBHu5yEL}X}bObQRQZrRW7wm}$gEMin>7{px z0*(m}4`fPRuIz~;UUf!=hp38yW5xP_0Hbe2RF8AZKp_@zBvs06e=|S+F5BZhcPO23 zs4h+&T^9As;*&%N&D9EbSf!=`A8!eB#IQ`ixSb-ay`FflxU41a$nV-w>gkClnp)a? zA)JG8%j_6Tmg6ORC)#a`B4OWWA0^22qUZ0266ys6dQ~7w-E*p@MVkSxh2@M@w|g5D zf-xay3&ecEXc?3hf$RYnGuY7_JUlo?hySJs*D8Je^(yUtSNev?t>AuU^7Z!PQ4O1L zdpBlW>XXO~Mjg@uw1|fs8|wi8-58Eyu$sH|A;C`}W3yGSJdh39esy^^xb-i-=& z1i76h5a8-OZ(AyE^TK_)>sidOjg{)Yp__)78+zWPE}k2pkNHprNOI>+`$n}M3RNax z*+VR3PZu`)CK4vl5S{Gnf4Uzk>hz2Le>6sLn!p-oECRBSE-Hfz%2tQT*p_xBmx#c~ zDn$DG{t1|oBvr>#-yaG7)IE~>3*8?mKaaH)qo1|OqO3;jduvq@M1QxOr%ir4UhA~V zfx|ToEF0H6&G1Te4)?yL6VUl{(mqS%>K$C`bGzp7Iw&LXc41^ZuDGRobZki1Qts!} z4e})e80v_X?uVvph9(*dokjn=73XL*okNd&IbF|%bwWnXzI2>rMxqCARM@98#LTTt*70;qt1R@3v*71#Ui4@Y-g!?ZhrE|4%YXA$Zkk3S=W~`2Z$!nE#_V| z)N613J>b|X)~hw)T2>`-TV8)tA@O)9fO6f^VYsTy~=2BM+|DuHQiIPlZ&} zPd~J?+e?$+r^dW@Q-L?A!Zjw6!Yj;M5o9ceo`N?EhCTacpkLY{?QhfkfXw7KW~jo% z49l)P#njDS=%2Ab(vvXmPwR%e&#!TPoR1gqbBjOm2UZR;?^t&|zTS@Fg8%jvk&RYe zNb+;nsa}_=jFZ*Po;8~E+4}=>`u$EhU$^u%B>Q-SAM%#somZ!*$a;M1 zO@)G#|JSzz6`|>;YT@n3L5xfnAn%FK`^ z;{kmJ%t>;QFzE21a(5d&TLsdKCxt;W)PTam3pV88NTRw!<$JeOl>K#SS1e{u~ z2jO0wL*I373M6AB`veo!rnqv&oU|h%@svmX{GK^@)}cWIkPM)z_=&D~afTPF zuyKB=4LdKi)csKRC9G5Znl9K{52i+VUu0fWG+E=^Atma(q_;)ku{lOb zEmn8;M?LK8^?Vu+^t5D+vQGM88hUV392t{qc$+&Qo4atI7_rtvTz=XI(DdPuKg=_= zURj9gVfX|Z**_BEsxtSkeIsy>K;mrQZ5UJN5+%WsGLe(lof$xJVUmG7&W{oSsY7{+ zTgw5H)0ct7dM}2`*1mPC{fn}OW}x)i?rbbZFi{~J^*gpV95p!WHvzXRn@gfbrGN9@ zRK?i3`A9E@Lb7fQ=udTCk%tnMI#zTl^<39|b2rh)MJ!5Fe!256!0{4p*CSsiiwmF_ z&g!wW1>6NTLqhXy7+{za`2HU}rtkH-CA9!*Bh4K#8RkG$Fg(pe{qH!Gk=Kd}Tv;|E zYDXWWmwK~b=gY{dTiJYZS`gpK2MNx|_n|qsXy>bP+=u%r+y%PPF!rcwtODiQv(G4; zp72+zDaWM|1vI#w)DT>bZ8s745N*;U=QS3m4nPD9L5`XCPyj^GxC)QDNe%L)2twsg zoP-}f|M7SHxftFb3ptg&r(513PfaZVp25@Zd1QW*k5;dn@Btc#b$)MJn)OH~$poX( zPkveP+aJN~#SZw{*q<>3~VeHnBvFTI^__Uor}`Wp56(HnO<} zj=_kBbvw(vxRyK5XD>uA>dqU7%bvqS&b_;kP5c~Z{Dnfb1=0G-t_VlSVe zO&c7eum_65VyeT9G&2Ma;>GF>oI;LPZ!31%ztvG<(}y|nea{Ff_msJ1eHt@jbcNU z-OD`l@%r6sKlejE`@h=K3!NV$_WsA<_X$7xLQGL*KV*{Q)HZv(4nA(f#(eO8^gjgR zRQda)8K3Ch>IesZ9G`KZyr5Cet-SNbeeXa{u5i_C3n@ixhgd8M3h)t;vN$|+XtPXl z-Co8EkcJF<_*OjG!2w2uC!kK%*dcri|#t8v!K94Pd8c{D$ZO*ok~ zG>U_Pt$68XYOpu~{v9#A{AKTAdjSFv4F%_Y06}+nn{R(VnqXinKCFQ#o6I($8hIim zRXa($QZ2K-m^r&#@Vvs=;nRCuo4tL(g>yE>5xDZ0*ZBZJEmCIzY^p9v_;D|&pU#&_ zqhCnLyR0CQ2*Mdr@sTh~a?kG@qGt#S;+BptoM{Uq#-tVbV+2E6U&(?sM&UrYvT`s{xyB^#9Stj?;sa+Y*gIqjc<`0cmvDo1?FvIlj;dGYnjrnD8 zNIS)+S^SC#$)y4eeNw@Ccu&yl+??TQpUfAi?z;Pi*X?s&4lJP?(3Ue2p6Q5uXBhap z1L1g;YS|W04vz$B$xDs>hFX89{Z3~S`p#c~+C;fz5XUl5aFkd(^E1O8qkSrT-;dAB z$j)5jYt*Z5DoiJ@(w~H?3&4zAE3)Katnw+_Kd;DvtzX+s$zF^~f|is8dvoAy3Kapt zf^NV7Q5tf{k2rizG#|zNkyU+9$>G)SGXs;*NDr|w$+_ttEXn5;;_rjSh}8CKUvJn^O@ze z>kqBn6XMfO7yT4wLt{c?d`{0qw;>Z2S^yg8iqj^asdeUz$1C#5`x@Zr<;47V6yZ}#0 z4tl#N>iPV{*%LQ1KY3N-4TQVe!);Q!0`^{&QC90E!wx6SbIaS*tzSyXF-CFS%d^c6 z!&~zJATPbU6DAfPFt%T-tpAFsl{2+E6I?1icj>RjKHF3cWRzkJ>yK~X832RpcR zpr67UsfV0x_&{>m6dbgS^M3wt9b-O$pNPUid<^k24GsGUV8E5BJGG`*b+ef5FPbLl zA=rfI1%#K+eAbQ%jir+h&Nof?h!n`xXzpJ{u{i>>u=#rJ~lN@Ky7<%RhQl$aY&Q%0412+RM=4X zoyb$Vxs5;ar=?t2WSrVysD6A^wB<>k2H@383A`7Vu|_7(elE(xw~rRdhtz70nmq;B}~;D4M~q+gaOv~4tu z8?qP5Is_NZtGc3v?n2$?P#>)sIX5y*h%3U(pK*?cN9T02t?|wNKRR5Yx%q%WC zeiptkOH@mGf7@D9;KWfxjM2TH8`ib?hM1Mj50ue8vD4dOT&&jo2?yR;{n+Q9d)r)m zp(TMTGXm?hFMq$Yg`Z@e>Q$Y_vKpZ>J3MfLIvd1g9pRDjI(jNZtn)~QnJ-<~ijnfF z%RJn|a)v03_QUbFm1k7D7m9oMpEtLcZKA=A zA8Vq4f7^WHrp3WFyhHo`yp!zo$7H^cRspY53VjTsI#Vw@S<}x2&v9OpEuN(PJU42g zI$RIAw(mj2*)J9H%&Z#uZC0CiaL?TChyxARNgU=#^Xl+)<)EzK_g*h6lF%GF_tVQM z45ttG^Y2;ozY)8eKVEh$SrIZNyzE*Da3Y@0l=?Q8q_Rpl|qa(MU+w#YJ zEsrpf3~?PWnJQW>{DuTzdv|M+ATvD(|a^Rj$Tph{T# z8lqfdy7cdNtQxmee(h)V&j}o|%ls6;0e_<@!HU`qp7kKQNe1t{yDI_l)#R`#M6Jc_ z2Fqitxkz&2PkFW+QZ!|hg(m-a>;ds*JeNFMTfJi6WbQZ#UI`-wQ9|u-@Z{>*7;jc2(Rt!!OJ8!!((bQmHDmN3(5<8+B0MJ+ zLK_Z+>vQ-9Arr|C5B9q_K7>`9FEI9d*JL;~WpdLN!#2*(;P32r?tl13cHc5L>zmru znE%)c2{cO`G0}bka#OU<8w_K=;svY z^>@w7)e`8|#;=H+a7%WTdxX=sM@CIVIQD`A&Or81s?X~ko*zr+Pu0{V5%Igr*C+K3 z4F4*QFB-pJ#<<)GO4!_9ar%qb2&0k96biw+mmMpkoRJ$7`E)4CmZAXb!v3R#Me#zHC-l%4J4X@Wjw(y@OYiy5|7oVBr0=s>6B*5T0{ntFu#{Ko#l(tOW zR+21#)^Y&GsZ{ex_>v^)OW`x;y3;D)qx!o{&nnTqB23C`_L=u@(ef%AS_to(cW2Ui zc-S*AP3K}K=Hcr)=&ox8H!+DH^cC5iik?Toek#mTF{IoDo2<;Pt(lMmm@UzgVElNy zdj=2105{L7ApW=pp8vvNQ<@0?Ct;F(8;yuhHayFe>JmZwj?8aSsL05h>kllk7^wm* zdKoHp(R?YoLs|RvbybPF6M1T~v3ANcg+-&!LR+xk#ap#_rfTl5z4i1JQv=R*GrHxu zK(4@WtzfAGo9hsNQq%Y1$c}>Zu$SUE4#Oi6`vfUd!QS8Wo*C_@-6347LGR?-&Dz&j z?SjKyaDJi(ufm1UU3}st3-w{Kx>xr@-}HKHz7Lx9jIN~q`pVKASNdE&j|-A*J{P30 zGum-1{HM11HDGBD1CRZ{^#Uc?(-mB{jqknwB%{a}cCnP~pPl?<$MJgZZp;T{p4YN_-uHB>8Ru}b z^J3%n2EoC!$2$Ji+BO>#+o|H?!*J9Nq~_oKL8aHM`odtk|D2YVu%^jw#lBy#=ZpP` z%!L4|YXE)g^$o-`G_W|UqG>1OZ&sSi0~-*Pf}=}qI?r|h-d(m>GEVL{Su1DC&`7dW zmRYXH9EG(9mraH&p!UPcKEw~CM|Vweeph?5L+8;m?^8kv(Ua0Z zN$o6z?1$s}m{KP9hXfPqcO@1g*yoG2-(I?5luBc7LkX2fhy&=5`UOhZN&Trq2JbTFc1*=%RSULBrxWmlRCv7Oar|W5-ZAI(yAw@&pS!8Om zBy$fH`f{72lHGSLv6l`V(^f3j9dfTV!}5y+i0|V@Io=ml5uI4X=K1&=4&z@xu}Tb6 z4$}b`V{R)`xfs2|C!~W%eCO=0KFaH6>gGtA`(C55>4y~xeg5u8-W!^;8Pq57v4vpU zw*233)5wctKI5F$lhZ_JNH-MbQ==T;WxF2)>x9dO>v%bb%X%@0Bz&qJ&Mtc;U?F!y zi0?Nfb)W<|pZex1H2pmTgv+w^!kOBa4zbAwIO>-i!x|eLq2GASO>dilUet$7-5;nC z%29aQpQet=m0~Re0`?tu# zSQzUeyw^ltzG?>`SGJv>@oaOjI`8EVVNmcWTwviop{V;CnIfD^J2o*hztC!En=!!q zE%*%4VbS0dKkA0kmEt(Y&c49k!@3=;e%&11U($y(>@V_PKsprS9{ zzh{AC?(|j+6B^xdm#)rMN#0gBa!QEFzzv8!m?1lA66Y`;9(a!Aqa4sHY=CQ`YhQeDJ2N~V$(u?hIsyTIPFpI)9}-l7{QdGyeavYF2?z^ zpF>C-vHc(}B6Y+X`7^(Hts4^3{Z9yEwR`<2^{qYe&Iuw!0Lqvb zLhRHlvfuVI;?+DH&l$f3QKmQbA=?riQ{0!ip3uJx*w5p+{E;=<92sVb^83hVQQ4w7 zJZvQDO!+@k~93!>|THrlD0KoQN(h&!6{onNUL2>8Zbuc7W8z3YOe-8&+oam!>mLCGu18 zEX9&ipf7dld7@EBL`pj0B?k|W`DStXjK3O(^Z4rAofAXu>yH<8uHLRpwzl=#G0nq)~V0t4Cx9$3&@g(lLyGz=f>IZuB!w;RjH! zP`DC(mT_sPfilv5-Q=sD`pep%%c6gimn&uv&cjAN3tmEY{VP!vDl{+#4C)GLnO*&; z$`O*O9PBFR#x&=deZN^^(9QawwbBuE0nOBB@trNw7cNup3o7EA^>#w9#tULb`))V6 zbzXY|oO`M~Jb=>1WQ` z_xdIICdipQV|aWNWLV+BPO>`Xb1eEp&pNLGzZ*1c9|F0`veA9}Aaeo8(u;4a2#{-? zMmz~9*%s|hw|D*1-xjuTF+ZD_Cv(mL9|Cy#KG`C=7C z3L9;;uOF_H&OGL>(qTdzGr@1%KK`DbcdCq$(#(a&PJ0_BT>J_s7l;NGx1j%_%G-o9 zI+)Dq-FzE~Yhv!J4t2Td%AthU^ISNAS;&#H^=5e5e=K@at^d5HX7inAg~zy5Wz@Z2 zE7pN??6PNg(%acum5@Y~=7(>-=*5VCtxLtCGXg3lMr>2i#--*o`AE?q1D`lD+GnIeCC{Y!;|gTCk!N? zUD>sr0~neWxXlM3z^Dh8{A2K!ZERjcwjhP)+?Eb{lr_r-N`;3l7R zaOd8_DQ?`o6g%!Ez+T_EIP`qTHU69&lK3)rJpR7HF%=D63% zh0|ACZ}mmjgQ&FqmMS+M;4`$SPMB3? z_c@qG829d1CM5t!u}mtFUze%p0B&F9vm~LgD2`3v#TbP`|KxrPI}%VdDy&Bj4L6;B zR7wM-KBWF>b>|P

^HD)o~TOX1^sj`*tuOHomE=8+n?(^0qR^q+4pbk+A3k)kToMt}p#$kThM}q{L&^vF8|1SO0pJ1Mkq3u&9l$cRsCIYo8C$+m z@3I*2*gd(J1LNin8N;29oSEmD&95WsL2iz7x9xjf;7CLba!k+3F~goja1+$}=u}yA z7JGD)B>7-81d_7$))JCprj9ECm!Id=_5>vZ>3AoO^F&(TtHmpYOKuQ;5>;QJQOmny zrAP)916^&V=;Lb@C%o8x?d{$OzS%j(=his=`Bhz3xg&l?{g>*npD@HLl@)Unr0?51 z1;m1`PbT<7FE|Tc(r10&5kZg2XiVA_UT3@xJnLo1ewI(^EFui%&N&f*MLkOjU^!$& zUs`tIk-Kj6J3<6tEhzh|hu6?(H^e7(Z}oWSFj|K;;S;&uG~=~j&cDl-7`7p}d6P_B zfT3Qis=7Pj^!a;B2!EnDd zKl#V|4D1yo8tC=@ZtxiM6N7H`D%|HS?`j?;nhh7U$%l`YlhYBpR{g`_r z*h8-n}bSkFE;*FQ||s+0M^ku z7Nh?qcO`YiHjRG{>%VM^|EJxR+&}Kh$4k5NkAK{i|F|pvaaZnJ>wkAw0(Xkz0{0k!xXZ3$|)N5Y%v>%4=RgJNYvS_Z?0Y1H5C>hm*;>$jOcJT9Uh!piz zw{gh+@7S84U%=%|JV<67kCVVD=il8+U8oY@w9=Eq;2GDE(58&nhYD%n^a!|+bv3? z20ISkm?I_S_hLlTSy%1Dh)$q*quf(CX26cpNNe~275HJ+d3`VG#JPN4d7r~;#uG4@DA1720%o5jOCs5K>Pc4KW0Z zS4;A+A*f!7LXz$V=G>WxG`N#Fkuv-ifFx?eq4!Xr1cy06ivt- z!?;Hx#33Sj$XPrET)b`Z+jzNX$c-iY`aCm>|8AMb#}w7$c<>8N=!Ma7r9(0C6|l+` za9wuX-w=!R9URc@Y55GSmX3nJNY3IT;asAdwa0QOuZbsRD8s?Ez~$xUL7uu-B?vZ* zm(f)ffyC|HF@AYQ0Y+&@Nsr}+5%*{V1tg2C7y36b!qHDk;`dVnYvSdU&8JhiVnkc1 zzu2Ao)E|@Ng@>bQq%MSojHHUbp!Sy=`{|L*`*~N#tB!D|*n*s-%ZO+a%J-q`B*$jQ zOLe>S=;Z%W*n#vw^$qQ_a?PM>0_YPTC-u4&{SxpH5vANTFPqRf-k{NihE|lir`otM zP@z_sF4jPttbdp67TG^QZXkjLGi9}$yP_*0jL1D;FX*E-iKy!oeitbPk5>2gd?&t! z#;{z^p4ZHevqkkrRPdgD+K}?`BQE}K*SyJEp-V(2@m}WmUg4f_auab z+Gvq1eE|3D8S@al!Gg2trv{~&sKMz)^~=feelO;ljM?{c@HaJ$HjIGdsTafOdM@GV zovUpyCrds5QU5Q1*l*(|_0v|*lhE78d)!j;IBR8OAp{N1&R%@`*W$q53Fq4vJr^Qu zNYPM1S!HK2nequ=Kdtu5#KzVAnIDK}qdPJVrA{XyE(qUMb8N&p2GrFP4)4uVg=#F% z#ov^mzM4&gxxl%oV8y#V)-Tct^U3+4`IofN|01jJ6|A4{SMKBHwAtuBj?sMwYVhv& z0VdtxrcW~%+#?OzpfCui@M`HEcBwm?@#{74&-&;|t3bp^x#+ScZQ8V7n$=3ZK=367 zF)cql8S$8|rMlQ)`E-|SM#%n9*Woch5&C5r!A(cBkv;`X44uNpA=EiUVm`ml>(FMO zy6I?!7=8l|)4U9L{3LiE^@HFc?UhFh#WwTx@YZ1+9mP$Q_Wn=5nk&o*Uh8E%Z#MoQ z^ts)!Uf?+C8uCcLGfaD~9d{}cV2k-l8JUya5K{h-uc56zYeWS*dl6J;r;OyRY7@Es zl>vSA{e2^6C*X;VAr$X@o_FWYXxJd9q{08wUv;fu2fj9VwdhZwJINsWiE_L8hh%^w z(FWq6o~_ws(eoOS+lR*n#Ube~+|mG6N>wxyE9a;cV*dOE%%4r_zT zT;JkN-Z`JoN76H)Sfj9I;(}In799~AGXsnsZ$iSn)(&~KqeJF(tnZ~X)D**d#UDV` z8`_i0JNJ3N=L5F0;PP)Lmoh|vYoEMT=4MvaljXE3K!Fw3#)$|89lC;=rHE*FM zZL|jNqd%^LSMzVn`@Hl?0jWzXDpR~)3~?kPFV{Z)7A>21x-MAuL@hdAdam~}1+L8d z*MCiom8FsHIwpvFpTMpdF}jZt@-fnKXz{;n`&muk;03p{2nO-+aKNjn~*;>P( zXres^JVSo`ga7qmw)xnoUF!hmDCQAPp~FERVu zO3;dnW#^xIi+2!@Pq=<{B;1p-(x2Y?#Fj+TdCfj4s{4zfv{ftG zX92E0H43jd9NJ$vy|lZh4WW36;8&)5wOOyug~ZJ7oA+38BCnod4+wO0Bt3hW#}J@@eiT? zLr|Btb>TQQ^M}}MHPc|GU7vU790dr_03f1-#YHv8&?#7~0DF6!`6`jg4lja*+|s7) zyWWQMHIk7EU?^{1(oQy~9l4t7)`xTdwpNHf8&9)R;pWK$Lz?A|k@TCort9=PTA3LI z6t`yxiZ}Iyu(=vziRWQzO5pIEOu2FgqE*))#)pVAef;T&I1zhKPf%C!HY+lNylx}) zDD6MlJCiO|v26?gB?_@@sKU;id?ny>#FZ{)alWS$clVW8sPYp*ru zH>1zCOlZokK>%5(`R99%lVpOp-wU2dxtk8hXS~<0zrlCJ!<{^~NIk#poIWgd`1*Cg zFzNfnDh{5R+apFJdhRWxqj&;Yy})`KZX!&`*Hj;fpx9Tm7Fi%R7+q@Q;enlDu9tp{)k<}q1tzy^5NSvR( zwH?nH${XJFZrbJdR@~;66> zf~2*}-sz}j_C5``HqF_92k3mPpx~0NBOg7S1GMb>wd(-kTyOJx?&!oXXbHj{>vmZ8OMKFQoJv@_x@wDrPeOL?O(l56-Cesq^4&gJ zmf`mvWPQL&IQ6NF< zz-s$_cm~SV1w;S$-4q*HKkatAp5oh){9*sOpSh_wnPRD`(`QKfgt8b@liJySKZ!3T zkYSZp-(|Beky>I=e|=ibtr3s9=mp5aUG^$|TL#FQ{yIisO06AS(T=&_BU5c%(aq)U zUEX#B-R1o;2#D_OuFSZphHC>bZl^PaU-Y=I+GEsip@Ncx*WQ(1GLP@eY-NyxH-jiV zW)#kl&T@DYRW*m%H!o;p`1b66+=aS=ry2}v1+jwM!JM?O5tE?X95f*2=a|FC>(BM< z_s`E~zZth<*rpA|DEzO+cE-n-CO+;rTx?BIb(n}fY7f(gP<5Yv2eG?uviKd`X&>yCz^Wqe zJ1p~eA}lxt`In2TQ1>DjY0xPUK6g*t3suJji3h06I-DcNht zTFixX)SH8@^RQ+sub~&>gD&nzRFeg!k5tI;YO>^-HAl1(}I%0lUcVHm9V|?JxHt&ENgS zI+A{H!9v(6c{lgF<=Cd2dz*bR$s!fCR*uAeMd8(r3@=&5FR(_Ry@~Rw1U^!Z79T}2 zm(sUkZeEf1B>r^id4KD_0S%m}<^@xL5FdYR)MYSOn$9nl>l46kM}u|vF5r#SRJm;K z=ROAe{X|$=n#B-M_T4}=f1jH6CT;3|(`%st3iKc&I0+nRXMWz#@aSaeTQKb3ckZ1& zroR)ppLJT=gRhZ7NueetXCqBsZqO}|L;y{W8Bojg(2w+UyVk@F#}aoc^@y&!&(l6l zmDu;;wSfDT;aASIO!`N~N($tX`*mJPqUQ)d_Q*l$P#v9iz**r!cFQMMv-Al-lhtrX ztVYm}lT+KcSM%OL@vTTdfhhC4-Dl;9It#$)U}#8WG<^ggHg8;20MZOU+|MA2Cf)D& zPG}8Tjmf&{f8OxGF!rq*lxpwQb_VV$e&U|#JXELTdHsoz=G+AsX9Ywi@hX54Y-ooq zM)r1CGN=OHP;b_u5ifZmfv&*ri4&J;=!FnEXzvnCKawajw z^=h4afaJXXd4P^;{3br)FL?O9Ht3Tlx_kSR9IayynVXN*eml=q3K1z~WYbx!+x?KI z%Y2a*eFKWb(R~EuNGhf9Jx+M+L*PvK4$AU6BV|fSTbLcGiy6O&sko!o`eYT|k7x@D zfQx+v(uK6Rw&DIcw6FWAao}EdD-ii(-JSHm=#_3gtagtv$DJXy?A)>^=tjuncP_s% zx=7fgS~b{F>_~CZ8|i((|NM%lLa(opxBEl9V2k!2g6gi#2D>Mi^B9v?4A$hj!I=g> zC~M)PF5`n}-X6~%a0j@M^G>lvhlVPhSM|C_%b5`j=Hb9KED*-Ym-bf|hu@g4*O&B( zFMDO&qH9~N%Y{)+NI9p&`ttoCq!SM}Ixq5ic-|l!OegbU18mXRXcJxlX~lCN`}>;D zn}Pc{%|K6?pr)p1nMOVJQFwAr4M3zWG1{pN{Nn>CwH)oe4_^OKyW=j+?rSW&a4-!(O}t$qCK(^x zEamDxwA>p!nfdY7$#|!YqKUHbMp_`=GsE#BByNr1E=7$bnjW}_B2V7J5RNLfs!l0ZRzjJ#n?fc>U-D}Dg&t-7W z(Q4__6B3^&s@iH%FFo|S?x*{nxAlT7uvL3M>mgmvmW<#1yuZHT3Fo2s`lt;2mv42e zNlPY|1?ViipgtT8*zX#6V3O;NxD+QsGn7SETx8tyYvF&@m221~*0rj}ki+3%)eM}5XdD{Tsm-IMJ?X6M6_45)$z@eY9}pwKSSFd+m@@Fq zz5Av+C;W5i+u((!FA4xyC%GVO1n3*`-~LW27o1e~_&H9zsb7!`B>g<@t#>Nzgk0At zJHj=}=OObQ9Lg`*P9)M8WK<_kOGJfvO;cR{SOV87n9;Gdl|d z?5_QNDxU>~R{xCDlL0qu;iFT~KVqc(o+D(P3Gf36dy8d1%&N(-ud7RcB-0>}d1wFH z106$tKhtd=2-QBoojY-|zcZ@!Mhl>>{c694+8?MdeZ=YPdzHwitcS?m8=>4~HP5V` zH;*6=gdxZ{Jya#QK>;7=7d86AH>q_c5@Smbb0-vxe(0%_R=ZYww)eFZJ+M96?$oy# z{}Y*kClw!W9uhlyU1g9UrkfFFIr~Lwm;E}4W-~v`2|dNzr>BuPa1`j)n=$8GDVlJ#4{BPk50T#d{ZYWD&!tS%uBujo(aENWFN!H@Sqdo=Gb zD`?4!q2AzTvo^g7S_`o=mv=f{&H?3mO^*bM;zBEbo80sFl%nEgJqfW~vm3YeY}~yg z*xGuem=&t=**BraPoZ;jgHn5edrD~{p(6z&a47ez^(5Sw2q; zD4%2GB7*~G*blU2)rK%cEz~X8?1zajuGwPWPge)FhVL6BHZ>&Tua##p_>}KKAN!l# zdS?n#h;^Rg-#*y5U%GV+h~8#`0`<{P)HaqC5k}QsI^y=WDBH2NfI;NJUOJFnamWb^ zILF$a%gAtVad&IgzXM#fnGGfRXJD^%WUA;L5?Hws?4I<{CMsZ+-jDb(efUKn^xe0z z@IPT~D@sL{(50!r@)`B~haeF{oTmAJs=- znG=NAXj_i0nkMXOCjw3;NAmR>dK95CS#HYj$(vc|*Cy46-0l0G222qRTqS}xO!U-~ z-+eP2Drj0CSe*(&jWBj}O{w zp|;^TkG8z?1nIj>1q;xw&I9~WnI03&2VDR%^f%5PQTt#>p6*uA=SDS@jVRxpOZ?Iz zq{fX&{U#7ZQz-z`6EC8Lk77fL6LFpLSo04UB@&`9y@U%8#zA&Q9f(sW-g2(PA8zib zj&EMfr_dGDZe4Tv2xe`>e29TL9bIewg1RN50sU*k@j+hSeqt09$g$c3zVL$>ltTTa zTH3rWzN`4;XFin0aSbm=0#zPzd^gdd@UZ~XSyg11%Y=QaO0YX1UWJ1_7F#&4lXS9q zy2&4KR31EBS7Gd1X^u$i@lp$v^lBa^S^;iCPkX-sm@7<_)tX-IiZaruE{-U`7*%*B z^!IgjmdUWAFFoviU6x%&BQAapTJ-#*q^vb$Ued|^9N?|SW5IU4AB?==j(hGDQ3r@6 z3ue-8KcC}`?fF=58DcugL~ZZZX7q!P1Ubmg-Q+{^lKgUMSN?>b?qU@KXvuh}`YFT| zNlV!z4f`8Kfpt6p>&>#rJR9_V=Y_m++&b)DRjKq!W4`@ir|)vP`%26yd5n8yb3{jT zC*WxIYofs?qVYBI@J*h`-X_aCk;2cn-$u3NTYcI;kr}Hj&nu2nkn+1hfDB8K#OMH* za4SIzPO22rbD4wWqqMT3yrdCX@OtzqPvFzq=B5c{kpz@&zMi*R_W3;?ccIvhVt4MT zkAPV|epz__BFwG%!XaYbe+M~v_@`{3uwqt7PXhMHwCDt6qm#5Z9;k#VDGeu|wZLU? zpktPY>8J~p?1xD16ffTh@zu!HxE|h=)DrT@ zJ@6F}Kh)+?!9m;96042BJzqht;*j5-gW(0pEMSRdDRW^_fPrH~wf(@SQMEqYsf(xg z1ckR(;LsUAiEUl4b~q=KCs?PTcM1Rx05qy7hkeOFTq_^%#4iXwq_10Mp>b&zj|V_5 zU9xbDyQu{T9X*sRW%i17V4@_*QsVu#GW!Q1FVpwa->MFheVRG%6>7%6z^Un@BQ#x|VV@I;vp;WHepG+B7e55B27hEG>;{i&H5S7U%WBD&BxT{w8o z_QzU35_UfhM-~sdqfnU7IFxenYz8at>|L1z(iQze(VyhV?{N0q)Rz@UQIM~EHt3c95B z38u!KbkE=HM53(?+kQ6&Pju^2wKo*b1AMV?^ zS-y|MWqE=x+eiKD=*U<9j3mJyBRE-FJx@M{Tdu0uj9YPfjxqv1#Ui-GWE-8up!Vvm z+@H=uecg-ZJ&;jL=85eSG(D%^7`O+WeoqBWPew zu-8T@RZ#7($ous$sP1B@;0$4Y@~&^`B_IZEQ4ggf?eEVMjv`$pSJ4jWvGy38idiwX zGCauI#$`{N%{%OZE0X z^_>wZOYOxt8qeb8IAIp2ftH4EbXw|TDXY=~Rrh$9^+f$n3QhW}PbcKs6%P?qv247N z4}3H5AkA=EoqHHqVCJAt>%bFd4J8KadP@RQGq1Em<*AjSj@x7_( zY2VQV^HRf@6kQ0~c~M(RH)veVYp=!S+nz*lWe3coQad|Y%-Bl~jk4g+IO4XEXf7rv z0#1RY=Wq|5eya=nMX6r$$R={WqAI*K9Izm8H&rOG`C9Y#CF)|k^fv^c5A5u|i^%uA zTgp?s&t#=hZTx+_uny9S%-iOxtMUfaDkga>~%OFX#l*}6@%txx&+O5 z0v;OsZo|9bUQ^h#xky@ic|7?|KA#RyJD0zlJt-6RrBBU7)ZSU{0E&K_`Pf%0y){aM z;m>gcc>W~N8N6jC6~}~6$+M4y78flX&4!BGs`|k|7P_}-{)r%FC$mH*K4*2W*eZ9U z7<3g$QBguNE`{#1*2nzHT4@^t=$5 zm{~o@){wOxgAbG~3ks8i7=5DD{)2mf)r`<-I=@K}QsGJK_L2wpWjzNBgfnGt3AWUK z*!!ZNm(#Vc8d;n8`nlKYP>+1qS06oU7V+*iN++C%!NpAMx|gi;$JH!^b7=M01ZmXH z4#v#5)*2Jn8`Ll5-0X8V&>xPjFMlpJ`mwN*@881m+T_s#=t5>nqM0(E2O#UY_MQYP zGQiiEAV&S#7k&eOVTr7+g_PC|DccUj?){TM+PDxc{TvbF{s zt5*%XE0#+@%B$hdVA{IzPlMke4`Qgl%u}(I#C8aO$*7y&GUsJokUm0o1Ek#Su!;D{ z^xpmmLyyEyFZsIo1k%XXIUaoJ!mqYMmeb$`8L=nHD;&yq{Mg%~D3;GM^6MW0bPGmbe4&wfwV+Xn&WkV<NfkOzN4XDUyCh29UjzGaort?eJVTpo+%r*!K96!%n=&@;wYgEc*T37Rr|+qaFg zxfI3E!w)jGHl$A%c?4$#XpVtUrQ%JC2f1r-mf{3;dCFB|>vnvT4tt+zt4H+)KF?6d zAGB&A*y8o_=icl2?0q+U13!9Maa*k3$PcR30J8zC=m>%72vf68X=fG0+0Le|njj}# z%gj^B36LY3L%FM%xE~%2D4g7$5kw|9r~M9}*&Iva_d~x;ejJ{RGd1T?Yak+eUH7&F z5y;20&?k_Vvu#~TFCDIt4luZY*6l{w;HvMxogUdcCqTNzNlaYm^G^ZBFl}F6h=YiJ zIniy2rb$~N4({XDrRQYKQTmMEpWdkDeg-{!%MsZdqG*B)ag~;ue^=nwXAM`v4RInW z<>`uv0X<1BhBUeqj0XfLtuBuUY22e1tk$u)jj|TqT1PdO@HF5wN(2yLAhrZl&F0kZ zhV)RZ!V&arx~Y6CZTwoX0(KR7R_A36 z_jg0L5=lD;@YB44Q+X%5QS7m9frY&pL;SJlOpGx%VPGSW89cf@1|SD}oaN_gKwW?u z<_ePyMLL-~1Vi;V{Aq!^Z@pMGE-L_!Q^Eaz%_WG12&|GVA4`&BMHk4a44EyzLXC-ougO0%1 z?;j7@Z}QXRGfx)91FpDnrp31c{AsKWN+a$C;>#I3ZJ5jA>S!+rUsD$5RIgTzLH_UQ zl6#6m*eM?{U-G8TIWx}}jpT$dIgHPbXC4n;ZCod#x&^S}2rbS3luMD8lQ`dHR)i0( z^(fIJ3+vkkPaO3RFrXZ`-oTK(vJPzne4_phz@bIE0M`Y@6J!_a@9x(&^!f}u06W^9 zn;t|btYr}m>e0R)_9^%o zacKA#|3nc2Kwf5RuqjK>>pz3f4yRS^eU<`cO*e&=?}Mg|qMWm5k|BCC9}u zTGH)y6Vby%1UsnGq=jkwjUVS!`B=b6Uo}JlwA+*!F$CKt+)On+g=?`U)`fmQVX6ob zI~i+AZyhVOgp$kwd11d@AN+yHg1b2GjU~!2m!2iT#;q}=`(j=~g5GeJVrp^G6^g3Y zg2IP`7jX-F&qp9fZ+@~d1TO`a*9$Wbg&4=T-wFFu3~guf(A6yq z1n-D>sm);_7ZDMk>GP*EwY1^)Pqwatw?Ygb9^^#K7wE+n3AqPJoczg%j_n7K=0nOy4_pL;x z{e7Su*Kg>GQW39mz;u{@cObW%yOe&|ebwUQT!GGP+q5!qO$-9i8DB@OMRiCRS6|ej znAn=IA*v|GhBy(U{~(m=6yTd`Pb_?EO#sGVn;%Tp=X|VF!JK^g6z}hIqEKOfq$h^R zuKnFcY-z{H7tHP-X}3myign=tc1US`DF~zIM~^p*XV)AR$%Ir7uoRb!_riC*H1MmC z36w+5eEeLum&9bO%dLnu_1T^y-?M|13WzxH1X*1IY+r`j9ZMO~e?IBA^SAnD`lScv zernzY*?iQ`!4xjaow!)Q)SjVB#z9-V5?w15@C>eL8>n zZ-@Kqac&JSPFRQ1gR-va8!gN46vY!{w8!5mY)|pF-G}^InDU{%$#c%YKt@hol$Ot*}$k6XfTWz`*Vc3#1(SFQW@z=YvyWiZCa*#KhF z>w8QcR72dkV^64CxT7(Phs5!1U{NQ1$epNNnd^BkQnCxEF1x;r@Jqzm)c0<$@}|SS z8aurse%0>&t1Y82TF)6lUT?F+UI5|_z;*v3c~Hsj5Y)`<@7)q0lR`A3GG=Lcg26xc zqKT`G(O&#b`Y|&bD zmcc)JvL;mCt1}A%hEX!B;pZfMkzw`|KPAE?VhtmfJx-I|F3M*gQsItN9#TuckE{6_ z;JB!j!&s$w*H^LegYTz&Q4rD-ES@T_j0z7NwI!)pzMpJ#QL^znxI?dEbCyqKD`6Ws zB{R^Op!Y;{g^$$-;nbgMUmCCTyf5{?E>gMo*YA@{%ctdlcy?!!!^*Yxs~_a$RL zd5m|BDC6~2R7+Mo`i(*s+yS#(F1lT%$qOfEKl6yXuD35}CRL1|7;S&4qy@2hT`ZIu zaiY%SmcEbo`xyubjL}gVAr<8KU1s}X5E~FqSG~;U*1&w}4C4&^-ry>Td8viZ&PT=( z*mBQu--pAhv@K8>!2xaSlFC~s9bbKW9;#6o!qVX5?A)Bw2_B!bdtC7ousa#BejA07 zVM2xr!_?x6dy$Vkz@2Y?OIvn_=!VxL_V84ARe8Lp;gyp$Ep2SbC>NM+pYQ#e zq=WidO~~W|)>dhb4?%4^s$Mhs!u>^~+;&PHJv2poiMUny%A;-YpkTz?W$D z1{jW`yaBowinnzCnB&|~%utc82|y2)=i5Z(`*QW3cpSmD{mFeep9nFcBTjYG3vFvq z((_84yRZ}tFGyHrR(rcmz%#(>3Cpmew9-t!4ZLRHi6S%tDc*%uhDaJZFq{tccU9I! zQ8voy>d!`loYnSqjU2)V3Lg~0&u0ogvJZDP|4MTBI($ZAf-|#to%MK-5sh@z*%}zs zjenW$yQ^Mf@)2V0d-w-O9&Z&Xy{F#Ii7!CCkSSlGmVHlhpZ;1JGS542`EDR~h;>uL zVC?z!`BbME@kbD0H&i~HV(~>DPu2y`pk0{YlZ(}-^iN2@e3=Hjs6~-AP<0s7wpTId zRMq;nkmVU(g->6Ru|1084Kzf7^p+_+CypmvBZ2y;{G+LiTM12kxKM0&s!GCLKZL5S zaS`#NKU^G%)3rXoiyYYWc<{VEuUOca+5BA*zxQT?_rZmUCb4>fDelqPFk9mhU{<@f zJl~Nf-ZKDA$In|``}=_w5tl|f&)f69U9C-6&|xcC1m<;j`tp;d>c^Hf!<~Ch+nX2r ziW|dWl6Ud?e3a6HUR@oY_w}+7O?NaWHIDL!Ec3U!G}1^a84uX>_yQk%&L+PA(ej0G zLO05IJ5B4l*AMSOUc51N%%CSpUn1c_)oNP)Ne^hK#W=_#01uHfcS$NZ*v(P9##Qot zu^0!>=9U-9Vrm^0elC)R+y2cc9rlK0*g|CbxxtN3xPy(huC@Ib7wYNjtr3>P)Itr^ zB6#-3>9>yG>cc^LT-j1AMl^C(Cc4t2#a1)|AnrIZ?+fFdyknc;PH*E>z4G4$+LFp? z`iA#k5oPl2i@Dbaj`?7rez^GlgS+fbO51m|&L}mQyTR-&-%Uoo!eBjhjzrF2=x6!# zMj2^7p-xf2O;&IFXA#Cl7Xm8BS0T$=Uy0J3o#fFLd&_9FwIbSI)89N7?*sq={&~|c zaKl8XGZB?;K5c96M}e_v`qyuZO|6UZex$@Y3km&KYW+%gphjySY-1-cJAQ_{7#Gcg z$+#jzGr3Y@c(6pV+MVy*QJ=xqMBVZy9r)n$dp~?ikxeOgKnbHe_sP$-lxUcY@|cr= zdWnanza+t`e;x1LnNUu$i5`OGwii5f9XsDOduHg?J%uYmzSI5iTSdax|Kby7CH|&@`*Y-Fx2I5B^`6#J%Sb*%^$-CX@D!@G7y^T0v*?|fwu(x}(W z6!TEQ9Ujd7LJJt8<6py3*1@c2Ul=Cre+Mf97+z^RePk-mZ3Nxpiz>DF0gfBNy``T@ z5m3%j$a74l_a1(X8V;35t=>kQ5jdd}KUa`_)agr8)}-+{!F4U7fZ|K~n1s{dearh6 zSgF532le8HN}smvdhNyT&^5_VQH$ue&i(vc_5}tFz&4hU^BvaGv-+es*hSSHenja2 zB_CvTM0v@8VQrpD^+qcLVUMrR_w?xl+!F8VJk3Je(B+py zYs7-btGyqM6Va@d3{;-qi9{Gwx|ri5cH(;UJQ z;(ZxzUBpp69(vZ_7>l8meL>Ed?`yc~s(oV3B}E%h3Yy1#I~djlsL9iZ`K06>XBN$u zo=iFO{dJODThV=J-`st>IGTHfSx04F#1^!W=`0%apM`z~koBe7dPkGW)zH6R>to^I z&-ynJaf$|>rBaod^HPZbIRx0NaR8(xWPbyD;S}q@ihmZJ!qrcSNi8_Ti*I>^()Y@Djyb{WclmZ!c<~HfWxg z13H5@3gcaTE0&hqn>v(db$tQ6ntyOrdz&#VP#Js`b$o1?PlD*GVhp3dCUeh$>aeKs zIu1vPepRENdPjRa#rEiqtrmnKG`(+CAnrFL-dr5p9~pSJdd=DhjKMC*uR#}ne_c*) zIoW=g6QB{pse^=NV7VK^f1%kpq`}eZq`X7~71e9?FQhB)`H=t$*>65v11ZKK7{mG9 zKONTiMzq@9q4~NbzFZX|-9T311$6yO#b6{53>M&7xS#nuGEerV!}3&-WI8rqYsG;W zDMql$U5K|}r^pO1wY5JU&#U)}`K&pc*E2z%Obaij&@0B)MMa3Lm?qZG)Z<_^k6%d; zrjO^&7$KPPH&MDouQ9MEllH%ln~IPJ#1|{Uve)XkY4OkexlWLXTXq_A0ARy21b-Am zNRX=6Y$vVn(iM$&Pj{D)X7VcTr4?DUg9iw|;P)3wUzohWG+Zv(TZrd{*oBgx)reKu zZW)$CQf|%Ba~ifr5mqn)rwB)(ec{-gaj|oHJ{v{ZcRcjRuT6ZlUBCBFl1{_OydSrC zZ({ft-*)#{hek{EO?91F9Qt`I#r;c<>bN^Usy1t6Gt?t@`vP(SN|rLibx>dT@(xq$ zZ-_Jy(9ncuqQW8!i{yUKEgczc5^IY0BM8?MmHZB#iywfTvlFcAS~B5Vy=GX_I6%it z`hF(@gjYk(7@QF^N}BFVg&=&OV^N~;it@?UoH2IID#X5TB1G3O*@-pW;~g?EZ4>P^ zhzx=|9CXb1BiJre*O0ziI8BIB^7qL@fw6OV?O^4lCVFs*=|5EDF5wceVk69xhaNpRLNhcus;P^Id~UM3%K9$=+nW$#r0h*7mnWUV@1uq#EJ$){P8GdfNKar~~; znIlD4BrnIe_+S)FFH46-=k;st2i%9>-6ej?DaY0xzaWT}_xa|?>Ic;d>~#0Xjpt}Y zbF?sdCn=p5KGf|}P~rvYp#K|xF9)$h2j?0D&DViM^B}Yc%9QW1%>`q#1!(Z4HzXTj zao$CRdgf(!qFIMz|2)^t`G_o^+bg}jC3Q-zN;1%0iAEb> zb|4)dp(yBGF*$xz(DxIc>BzljuUgom$9(pFiu8vcNmvZ9@SA6(_*TMA5x&kE-VsMc z?13Wt-TMFu#GUQ?i;`7F$q;A0;2gjZMKUt-HqxOGjf>O%!BkGulU-5))5$+$h!@u#W8L8xHje zhN&+_@`KdM;0_8|Js6Ld?W&*IkUwQ$X6t|(g0=GT{kxT&0ftuY%RrW6nNq2*uuCsC zx9X$sMbLhOl3nzw1JE<9yD-O6e}+T&baK^vyGey@BF5>0&{=<^$C>RAz=N_1F_ilc z`;`n1LP(UhWGfk_5^Mfv*L(Ds4p3Mq+9>;#BNAJm2>&*Zn}FsjW#P=ihvpvx!ACw+ zc~qXMVm?kcfE<@`6Sk-6i)b?=Vt(p$dAncB#jS?7-mtFsD!lJ0wHTisy!F^4HQD2& zb30}dl7Ye;I*CX7$v7mpX7PP}iH~)f zVE)Xym5$PZzL7x1UZRMQt>ujCKpX8e2@2%N{I9n2+<*;V4o_Y-lAKUnr72$p5> zMN#t^AIc1%*Xax=j%l;k2UEME^ca-o^~RvlJ(DwgfZY2Sc4rh+alTCO6CLB=NHoGm zDmZ~lgK210W+$zo?aK)Y6qjiW2FIqD*@|`pyrv`Ox|DLSq*e0qLZ({r0Z9J)VePxP z6(^eb3a|J&B7U=sSv0GT4b>~;GeOI1{5}Z;Jg1k}gR&MdRpmJGA0D_9HoU-vmIc{8 zzMo5v5RI*24%PE5eNagTD`2+~0;o$EoF#Y_4xZ&YNQPh8!z;MKxUw%qMCe5=NN2raVa^9*XMW8JrGr_e&O1h*c48!I%e$w;qU4lw|-?jcfiopF<2kz)m&g$yp!&qnFDe+wOiL%@n$52jwP@bEd- zojuufaKXztZRvBhsmsK&-$MWe+NDDtX!$EDcgZ{^3QjfjRylWzg%Kf<>7v}r;BZ*4 zDKb=?kEkU+V;|JLR|YlO0~gbizD)@L&O7W0LN)uE(q)Lqlq}o7X?|h{=8=clbNq6D z9EQW@YNO5r*;wiG^+`8|b^hTLiC|yDhJn1`tB5lR~;ufJ}U&kboX-meY+f= z%a$8JKXeFd*R$P=uy-0oj$bhc;I_-~hk3W=!KUB|%{R;al zDDpC>m3CQNr=GvXY1CAwrBMztf7h-$+ZV}Ch+lXSA`%9mHT$|4nv}>E#Fu_OP|amF znLZmUw-L7Ux9jRhUmT3WGXxU`!J>>@)QwDoIF3pRNmdBfs$h)hTC~DG)6ihl;kXYESNj>hn6(u_u8!@Cj+!9 zI=!cS5jX?}Pam-ZTBp*N`rh(dJ__H96H$eD>Ny&tfTqAk%>kltU*~PGfIZ{|K>*Kp zZ%mAbt9@S!_1vz}+ojQ`f0F_aogqTnBuEAw0+)eZxUOA3SSl1FqrapS5{)d;IH%Ui z=3Da_Y-kJ_lZe#u!bYw3aV!v&37vVhan>OT7rI#<_%pl%49X73WWO*o7mRdaXh;X6 z0mLc7G2*u}IRHew3lD7W$f#+cdIK>y@$o=|iTXXgeJSU!@yX=FA-ot>EM%qcz!4@N zgYNj&k=4!G?{Yg`+sGug{(TCHbKcO!CJPeY`8MruDvrmm_{k{c(Q&%<%eH%MH&us2*O_Pzd%?y9{Nu_fZpRMJqwTmk!=S(FL^9q104NoaN2i& z^qKthw9QYP#5jw;-We)m!2QASn@}Ec$$0?c85{u;Esk9HftmA>25-rXJ@`6+y7)1C z6-gtI6~3cqSR!4VjW0&3c~4BZ#L3rQOWEVj5FiKOXBm{*g))KH-P6xGG|_xo3+$^-o;Pz=dp&P z*J+q|e_H6{`Zy5I!-Y_75luh=UNjt&n*KAr>|_3UZGbgJ+B^Pm&}%rh9sN>lrs{q` zr+l_#)jOVDc;U!z{Q8ctMRtfaO6fYnfb;V!5c|WBVgLZwCdlH4+M6e8^Ib08^YjQg z_W0DGs6klT8?tmCM`buyuR6^%$c@`aab0+ZRPuye?d|4#dqHP}3^_7Dvlyg(qz~CJ zYLR;KL>U3H_V+X6HW@v27oz>?&q-LXnOFHEpW;(G5tHxr6&Db^aI@$kHpb>Cn&&G_lZK@%G70Y^(FJ1yf*QDf=4l z{YOkPh)%Jg{~M%x8YQ`eXu=;Xgt$T=Xgm#AOZpz4_xCg!as&B|0urTXBO-{Ym3R!r zjW`Dd$iXJ_5PW8)f8ay6)cE~+MTH6G{l=aZu1s*y)BT{wcEFVg$5M z=qK!e5u}O&6<1kze=KE=cuIW7%3H8BGQ&Z7rP4%{X!pOi+|=Ut`Km|SrNj5sx-9ED z$4~1yg_JF!A3OXI{XbmPm4l7F^SeChREa}}J}5Y7bs$G?ogWZ0TSOXBLtJk*+|{4d z@}&atO)Mh=p@PS#Jn_WEwDfsUMgX5bre-0wFhmk(Y)AGOkxgHz{}pj}czpF~FVSH) z_sLW1e;0FhzDsJMzQtF;;Z!ESnSaIS*u?;UU!K|-5^`Hb;U2wgZ67eIgLgUh$0Dd7p2|M*K-g;r z6?bvVO2T_M?)tu65fYuFi=8kQ z%-$g6c$6)`5*hGuAa4j{Ek+xx`!?5K_rbke?~|QSM$8ZIX7f~!VG#)kDgZ0b0I8gp zPfBn1ob@t+JKcyLV*T@edABQ?Slk=TF9?6sIBsG2+!u2EC4XjxaqVkH{cET zVr%$}wLr4aWQHJzTWg9w^DlJ^>T!{Ysy3oZw$M5ebCZ3*=-6 zsBh|Pj5`gO&8hXxg4B76u?}wHfxCTJ8c1ms*a>iNJ~cu+1}I&JI=xX}({UX2t}zd{ z84RYki$S7hFQ~L4FTe^rQ({Hw%)p&jLxRQiaoTPVW}i}xE|W2>khCC6C(C|+2N?ny zxRN9_zs;SQ=PnQT0&xD<16AN91f)!&T>UnJ!?uSF&wy2N>_@aA?e^EFGlCFOMBo}D zNBC9l#}=rD&zoxyO*}!WJz|po@>J6XOr*>oYc>`=*|l7Ev*vkGzJJ`1`Q|NLyiYiI z;7`}_bbE1B1&XBF@0Yk&r;xk$ijF_TKirw>o!Q1`>5}~DD-4+)A3$;N6&HO&tsk$e zUY3Z@^i1t#567!aDSuZ+0s!a8`%VddOn_pKBce%7{pZ~(C);uQCiF7@VLn=iqr*y(#5G z#*~U};^FxPrnDV&6}TR@JBTl|kvQc2mhSjV>~4NEuWLr1-|F6%x#>Qtm(fBQ=}tk< z-GwJIqfa6-3s})*csW@mSg2OGb&vgI~^4WwMKgxEA4d^VLq7b9P?^D^61Ljl|oC6zwV> z4gsRR*E05mx|1RnW%su~d64IfV_G78rml_q9ShHRz+0fMrNFI{y(>216x{at~tl}29^OJp7uTA=5;{m4?ViC z)F^@Ni;mEq0V`BU+o2A3rKBIB*RYooIDPW((gpMMumOXWt>km5<1&Uw&TR;}|CtC? z3B)6`lXzCpWq3ar`PXHp@3-SUl=sX1#aftZQ7mU+49_O->L^^)8z$WQ&0@D{b%EeC z`xuI$!cgvt^r!sdcz?Gs49t=7VOhTvhu0unll`DD-4B8n?dEaU>Ae(Za6a?pN#O-Iyf(YLM>^VbLLNZj7S`O@~y{KeXP z(Oun-N{FFMkF=|rpRyXm$Yub%CdwY;?GsLVc8u!WAiB}L@It+a}>_(Y<+&X&4iC@R~8g~ zsgoizT1KD}b3EW~0iOm9m;EVU#)fEkxtS7@ne_Z!dhfG2HR`qAOCM)1YlaID10_5x z8*f|HzHI!3nxCQlZNPVD5Z(H~EgEpMwL(qgG6vXzmgYZMxA{pjn9?y^HUcr8!u7y4 zSzi~<@wdy1csl-`QeOkeImts@Gy;Z&XJb{b?<;c1+q%9zKf{CX%aHd>-V=Fxfk67% z*7?5heFAQ^bzB{Mbi0`bodRa7cpsX$i0tN-KftC3=5U9c>XQ7Pse)?BAK(5F(&P$> z0M`vFZwg0I)EmHK_tj88Qe?Ww4h|LH>svdZd4nGk0BRpQPG6NzFT;1+(yH!BO2Uod zke#VLva99Eiq;HPO~TJI4fSZN%T`PZuApNJW8Eecio@_Xov z+9_ETVCBV^4?%`$G!g+;P?7vsaspwtD@?30e4@788G%{j7t#{kaN3aZGQNF0IRJEe z==`V0NF@2g%g4IXTnu+_iXqiWo)G2A!S^sB-)g&bmosp)d5Nc~6f0T+`+eu%-{-3M zOjH}$eFI%x=Py<4CF&29)e)V~^2UEx?^PX>={2B`)08k`qQos%`fOI{+8iXzP#C5j zx>aZj{j6=ORaN9Yqf_i}_!05wogFrpM~vUdBpVZ`C)^D-JAdsrHU%d1=kkT|(v^`? zwc%+z#7Ca9P4?w+FpzL^M;0;_JeGxX$6M`W@8Pr~FuTRtEdAI+?!qYR?E`U)_+yd1 z^>T3vl~{pTDJG`?%L)zfx$#T{iv|zr`^d1Aa!ZdAn8nCvc)8*ENa(-w`4U~0{kK6D zV^(FFOXBtgL*HczXH2i3^1y!UpE7HmSE=38GGrk-jl@s^^@;XnP}`8mSfIegm|W&d z8R4Nc2RoTf^ReZ{8IZ-hvp`op9|b4(O6hD6Sg>%nIhPmT@yEQ5srF$FwcFOL(mj&a z^xY%mL@H5qNe6d9p?}XIz+JMmU(Z3)fo*piO6__-9?SjYj!J|xUdO`*Vrl&yn2Kr4 zZeS$wiJI|C{|?t=bmAAfZ{cEFg7?};V?$~TE4A0n9TwqI+I@%Eb+bRi{9qDw#TK8~ zRiyH|pDc;1?kw(UM($WS@@Lm}KiSCA-d>XGexY@|-JL(P(dr^?eX{V{V?*wJ`6V01==AE$wC+XVmLpudoVV!X%$S^6Ne)DlU2F{ zXM_-^tL$H>mH&y!Sw2+z6ZMtX@3=JMCquCAqYmlqaEd?bErp)7QS^f<21yCIuKc4| zTns!Y?SvY?Y)6~wb)y}nWq9P2w~j|8&NL_t__y;l7Oe4mz+~X}Sn)h75UDFJ#z(4AP!gK61$74gwSmu4Ca5ic|6qxub_U6_8=lMl5sJb&b^p*W?yv#; zA0LABp1_-hGVZ_shrTq6`204oG2%H9|9V80zyt1&`9C1db${`H?-oK*Txse&wsc%Xr~Q`XrZAO#3%STt==7|y4mGh ze>;4C_CzL1y`y#h{Rro}c<;oW;rqi!;~rb<{wsPF%QM3V{(j*3-ybFZ{gM5l;vxQD z^!cA>1a0we?{DBv7xs8$0ijAJxDj{jedf>R8UGJ|p7%wFn)&PbjC1=t6?yY5@pF2= z{3p#-_~&gBPG?ib`(s+XCFiPp6K|xD^D9UOSG3>an=BJpLoLsS)}m6W`-+GJMaa`o zw7&JdNWY5x#jP2_q+cCO!V;`1lWq6scmjM6BAdSlBxNVh?pev7>HfAQ-BZ3xAC9|LlOqci@dT4HGn?RsyDhS=qnRuV{Y zcjo=H8D*SjoR}{IrLc}u6NeQMhsr(l;efflf8BhCifNh9ECvz;3RCutW>m4i9+R{L zCY03l;~0p`+&+%as%^Sgh`(yJ`a5!zSWiYJ^J?(1&FLj}Z!ly*^AUwR?ro~&sjV5| z6E8otT@}9xNeX#Y|2h2Xy}Mf@XaaeG)UOV+5mxn18=#Xbk+3E?&2X15?1bE2H~Ee2 z*i~@Hn4*UtaryRdJsIV~A=(yDpv@oT!-#x`j3H2FQ$7X>=jz81$L%V?go7~p7GafF z28|+82nk-kJO_DkBp8|8e8$38^=pcKMzCP^_GEv_x(~!FvPt&?&6T5G0L4@@q)Jd{6CDzN5#d72 z%Qqvo>v>Kuzab)ZW*E4;enYG4a~5DZnlRSM=*8>aJz`mY=`%XEI#pY)@r-T+3?ib1 zC({6e39euVdU^Cn?JYD165LgmMq^MKg{nc1iF;q-VDXrFI7!2Me9k9F?M+f}rlEy(v+VusK*tM%dvn5g^wAKY0=YW8H>sCjV#Cw)hgDr1(b`YFwOG+7 zE(~q~@CY_?%W#wwkmdWP8Vq;I2V#}=ncut%x|2fodBuj%A{V41n77TWM6d@U?6X%;qRKhDL2b#yeD#3zu#q|PDTg#%jf*9Tf5NsBk7;e$nTX7 zvjM|97GE(^*9$G5b+0v$sCNt74k?cZ=$XF-@{)b#4M2$>s3pH^NZ?C=zAxU`58E=z zd<`p|f0OMY9UnQ!zde36D;Gk4`{kNxzXGFoADe#Oh#yRj6mLMeIdE!T(b$xK_ zwclH8$4+Hl_5ktazTEoHqp%ie7K80h^#=Dn#okFC(e*0Ag@Fp@w_prsU*_R81^1NG z3q$rX;oyf@FJily4h4@x+wYPN%Tf}*wK+jZSPy-nMfz5(Cfhmu--&LL8i7K-JDWo6c zds|y%5IocugMkd>&CkxQlTZ={u99+Nxjq!gO+UB2&8y7N8|;0_0@0C!clIOZEwa~A zK_>+{j}39u0X#jN2y(vS-a$JmlX5b?M_)j88J+GR4OX?Fe|AsFKxoHu759mf9IYyj z+rn|r4H!@^@%tn|@JvJBu3E{>e|GAn+Ys9pk7P^cAe6%kZ@dl&Z0tI)NcZ)NioF5f zOV`+>KyroqV*H4Hd4B+Gy$45Y|cnQn4_4v0}Ud6RP7jsC+Jn_1}<#C8u(K}|(N0LwN z_ql2F7NA+98K=5!kyW>hhTvVbsamKu6CV0E(CzdpBvh|*#?0|#9PX+f4sxvzP6^uM>mct>C&C;C-*q^7 zSuoKfjQB%)x$qlzMZ5N>#m}HgKZOoq)(aXfPX1(duna&!ksUO7y}6+;UySnjh4Sqd z@qdnjb1|U-jY0LHjAz~fh49!I&-oqx2;36tdip`F_>j562GIq%zgCx6 zKQ>DDb)Yc;Op19#InwYxYnCI=<9CxjsXz0!&u9n9KRi$Lb9)GNKW1%5Zoz=44m(be z@U5($`y*oz0XhZLHar_(s%mxzYGkM!5LE$C#DB60dAD(-Y zCKTuCt=4x!O}2v%c<)|iM0j-5s#1;!13sV~sFxOi1NbCA*%N+ELj?Ouc=wc#4k2|% z9N3$-#o9xf57Gk|Cb;-duX2NnJH7bE7wtCs0E|m?XpX@*b=n)7kF~AkRr!H$HYEU) zI0@42^@yF?)SCl_S9eJ6erJ13%MZxF{Zl&sNKw~*xNVQht0NJf#n635|I(xDEC2}Q zl)jqp6lsaSq@eASg_iU?2wGH5@cVjQ%an=2LDhsL_%Ibagan(&P)f_2REfId<2f^`}`q>Sp6J5q***Ij1i(| zo!!D;QYAM8J4Gf;>I9v0&PyXNqHh_rYSH~}T}P%1fdRD-Nx%Y1H%M)?gXU%z9?b(x z^jPwMvVZ1O|yiwF5U0w6QOw=jMH)e6XC1$(#< zWb16cveM0Vu>vlx=LfWEimGEX7un*wh&et8go#F`wSd%*#Ci*eb38b^B#_i%k+cq! z_(F^QfD=5X}iQn7@WjkEoUR4kgK* z`#4-mm{_KtH@SH%(lNE-VfO_)E}&$=;=ozE@lVhvU+>xozJ3HEd`$PNPyFEMks|7eDU@TawKC2d}B(NpVDj5#}!3&O)t0PmZN+}FtNXP3QETC zy2+(S84r8%;jT>6mc|3{$UP68i|QZu!|;i`a~|<|uZ-i|-;KKT<8|2@45NwzrMJ_A zER}t`rbPC+k80rl=`=OQUt9|RX>169(g^dM_QkF63?QM!r8G{y&|DKf!|C)|5R>w$ z-YN#HtZzO;tj+v1Ds`!0Bjjt1{vGR{SRxyF$m{8tm5@DvIfBo1LDt0V<}$Cwhy zS5nf98iB#J+7~@FN3c(%i7#PYz?WC|sg7y546=&~IVyvs$bFX{thGHgv4K7ueNpWz zI(yGr3nKu&ey0J|U;s2lC6Y5*5b-TMuQB^u;tZ>H2NU3W$t(ntA}lqe1yV@9riiu) z)Q5*91!HS^!o7ZHwtS}fu49vysk;88BU9C}PPYQ-ks?qrs?xuBqm3D$b(s&T5LJn& zi(@sCX7_sJaJMdySJ*Ov`$xqfF{v5qT-^^>DIR?y*0V?8a4qP&csWQeeSF&dupiNB z1P0AT_t*@#-~hHeWrQs-wjbtrfV?MXkHuT2uS$B=FEj2MmTV4~W*}g6gKBs|VZ+lz z_m(+uGndat;=Z?k)JHrJllAdlyNH>qtqGyZ`We4CA(g`CDL&Q<>s~?r-LD|e0#^yf z)m8Be=N6xA`cm{Ia2bedN77YR8`qG5QL?+&k<6v>Neb!esvVb}vDdG@>?_5hMIPDG zb>LyVtQuPSUayiPqY@|s9TMV;Hc+r{ti`L#Rt=ACjP8AsL_-SYQUMDz+4>sKv4+3ejTJ49y0mefdZPg z%dfSNe5u$Okv+cp$6jfx=I<7UPgoiS^(rNkNK1&Sb2kebj{dIRBJOy6;o`jl_56Z~ z>-=nVphO{b#YV(nc@~5r`CA^zh2ODjyf;j^Igtv{^ua;AE&iHV$c9s>lYO2-s+4Dy zp)*`JESfB6dWAhAAFr58pF8_H4vroVJv>kBirDPo5Tcx2HKV`0D=a#@ESP7@&UM#g zH#^Jw>?H$Z+^PK{614a6S;qrq)uKNB_8a-#Y|3`AQDySyunf@pAo166MywvIpYOt| zW=*$`ATAj51~J0BLF)|BA7jXx_z>Bu-tdyZc74_jH!aaWOgDT(Z!o(h>6 z1i+S;UwShqfs;=9#bq)G6wdJSkExmw(Jbhl~K`my-3b2N~eYk4zlixCA zzSOmWI~|RJxWB`=3_WH7hS?TGd#uV{v7##A*$VbE{JpM8Dcfs*D3yv|9UrXJC(8XZ z*AuK6;-}ue%qPesWdcQ0-i?qfiyO#oP2_-{7~xjKQSaQq=c5@O({v*q*%4`g?N8_4 zVR$;Q>;p{Q4$`S@p`Ow2l94L9YMDKuUV}?q@zxf8Qs{(^E-cK&fX0HgnPfyg8(P`m zO6rb#aZ_(6%YU?x{D2kYoWYb>Jp-=<-XiJ1wl?*6E4Qu66!ew9mh9Eoo_WtsLkVud z*K%L{TQ$Z91Ego7_qdUceIAjRBL@Cm#NkkzA{oVQPLUUzN>1J33V2J$mJ8f3=_Nk0 zynRLf3(Z_3em2X{hH_1 zK%nhBJU0GD*XPj#2Q&G3pK0VL(WoBp58AQB@HGkU@X==K2>h+=xf~JRt~9*+ISqF9E2g1>Vt*M%!P1(;1 z4{0z#SHFEJZ!5Hbeo=IGhFusj@(}IfYI)*1K0wf7-k1wxAjjyU^!;hRS$>5+h2Y-m z5+zuCa*k&THaqGN+zN#9VX<@P4bg#Fe$tTy8H`BNooMd!MFwX=4Mkyf`w8u5oY;0v zSW#hL`D?meF+lGUAC+Tq;ra2ba}BbWA*#ngX74YASx=S(tUe!|RuAod)1QrpV^}%e z-K!`PF3JXGBJgC_$!iZX@}#aOMgbYzMTvA8Z~JCUf;p12TWcaCbF&iojN>no!Ff;X zPH22T^!T*@zW2_yn{GLRDEDWeLrJ~;Zb!K4$f`?4|Gbx(+;`b~zG6zMB>8yLeG=p%mHPqyeLJbJ!~bpP#LzQg(4a1eU^d@eu2X zVk-J#Pj`m>1!uq}EJVzow9oadR8`yKbi?AJgM>|9O?4~_3NRen_SB3FGd}Q~_7Mpk z{iPpTed+P*Ha?;Wtr!B_?#oSZ@+@bKl$`7BwTi}ezpe%2y-aLr1Y|=f%t6n<{QPs5a2UKD%5+GP+1eNZ zI#fh59_ps>M3-~cD@P(Ye8lV(+DQL0w9)pgi!bJ8iCF9x^jT}X@5d04=x}VebK`JG zx)l-1giyQ5TS>2w4?4tEec$T2Rk8`1vq^6S!NL4#Wh&jxHm6tvohhLVgo9=rN4BgU zC0H6+E&M5SS0PQHwH*~m>OkrpO!Bf#9bUFCd2+LM!fm&13e#Q!HWm@S>cBjR(c@~1 zJZffIrRxN7wx#&`m0Fs-2I<~%C6UAJ98)QdG_0?_iahZ_qFeSz>q|HEsY$B3@%^3 zedaIpbMXA`Or5O9wFS7=Po6lz zMSrXu*ZUsw#6o;R99#wPo(|(h2Rjsxy{q?lDT4ihXxE(|4ecFZ!c>S2H3FS9^RG0> z1+@(?pjlFfqBZh`s*pL*y_Hee?^?iF_kM!_%+pMB&WS`G{iErjy_~{Xdj9nRHKUS( z;(JLiI66{gtm#WyR$^SK@59lnhm_;VHh&(P*<@bfG?3ce3iZd{n^~^fRZ9j0FYc=< zw8IbY&M>xhq2!l>tPlDqZ|iIAFl6TJ%dxvfhb!!6WCr%4Z2fMXjA2N3)h)l5(+*DZ z4>DgZ9%eNwk7RfyN0H|G%=&|Lh+xQr`PF5lrl2*2+u)Whv-)x{lg8!7morXQJuEUzr}0h~)b(tbBH zh2kC|U4AoF>EfkUs&{z(sE<7HwIDPq zFgh1r_q~Qar>_KmsKp8YZC2%g@IGzp)OnDHD)F5Z`h5MR!- z(+^p@oS5BPcJHN!umX#|H{UN10^Z049p~7rB@?pz+xfIN*W*12Ss1qmpNrx}C1-r> z>)3^D_VfzpC?F6%JE99+$5`1VYZ{xjqCj;DSvPwZf$v*z`}qYR!GUDWp#DFtmG zj;!3b*!%_Z1=6)M1Dq!?hK}skM%K?r_+D9k+cDAc4WT|o@P=@)IdJo`(Wmx6u@h4b zz`VVUH*V_>NZ$14s;FreHy<*q--uaxsp+x#rf5X{oW7{jB{})&`DpLPcuWq-D!~Xb ze7x>Apu+WFzYceA1ABGodVHLm3f3x`;9zY+rFqI`uzthQ!XOX{lv*Me&55$&H5z0}9lD#`t>A6>cCwv)^{B%aHFWK9 zG_Br_?{MOBEY={EoB{}CaSF>-AIvY{&Of+@ctR?V@md8^c=8Bi-F`;|g1kGt)}HUv z@~bef`PZt;fC-z|gE;w|Fplq$Z~-nU%rxL^o&q%dMEDI`Y`RydX+j2#Z;VgA12T!* zVsGEF4TaeF8VHCBY8667qWkAJ$17ecf8XzWK9KU?l|&a06B-4KiN;d^C`@nFsqKcJ zd%vt0MS*s4r8Hdh{3nhkzQ&y33JUE?_~CG4A`7tg8rHH`e|q%Y$K&L09NbEKuBrlR zn<9Fvl7w9WM1U3@$If;F#aDec{HpaAtyn%B+Z6@Yv%i%UT}NO(S$l za_`F#0izA@VfdO+;T~|t?dJxf!<*Ro&GmMGM&kIrz1i^3K{})cGvK7aFBG|Cj+t9J zay{Or$ru68bE)^cG9AuABW0*%8KK+Dr_*Kae<-r^7BLG&qlzFAXM8@i<>ZrfS1e_G zt{5{}UmX1$V{$(z*BK}pQ-b^#NaG_O{%y0AK7w8WP_Y=v;3`5B10H~Z%J=dI!oYJ< zvgdn20%WMGP5XsW5*8@rczm>MwNs#T5!Q@G41HsN%A)=-31L+x#R&4A6#F3$4L;M^N?#(6UE|RD zx?+qY+WX2BO^AO2iD33?q_QqA&3MeqzHf1RU;99M@^4G1#d}TXaNpHT7{VSGwhYmp zDy6RZy}%a=2_XIv`~LQN0sN7O19tiw(FGYUtv{i*iFd~Ni{hq0wurDl~ln6cgZ8`ed6~3E3A_l0!S53_r_;qwY+Dr3?^Jot6Vl*d@$c_e z^vY?nsVE5f=gc)@i)c(%l{M}!c?v17sufXX>j0?uERNG8G% zr8mvo*-@nAe#gEGw=a#Zj*$B~J=7?vzHMS&tTkScukwH%U=mf{$8Rh+>ZhdPl51(` zI#PF->;zfOo55Vs6GwkO#3W$aT6j-=b>Fl-xpG2Lo+AV+U8+9gvA7xosBt)F3b^0+ z1c0l|yySbGZQ|cAAehpbRG=VHFT~)|x&1n!mI*-a!>8+AScmv*9|<>z3sueYeu24H zp5MD~AB^}Y0vYX_x*rfd>u#hNbH7_pIzIJ3z_FL+n7KVA_bz8h20^@-`569ULolOx zwfaC&4XxL4_RuYKj*NY`{i{!e^9$FHa{Q6x=gGzRaK4uHYI}zpyYH9&z@TTS{v3i( z3+8<+oN&=GvOC4WI@Md#37kUCmD7LebNco~kO3}woxpNC*n&(tkE%Q)#s_0t_RQVL zhfJ@%^3V-F0?i0N?LB@eRzEo5WVeV;*Ntb_%k-wk(c>cQnkpRXbn5-v|Mt7@_4$16 z1pk0|-8$H>Z~f~PH|!3ePPPOCv-v_m?Mb&TdUf!{osS!iw68?Y&#P%#PFHasuv=Ex90 zt9@#~?DZu>Qyf4g2E<($@byg@sfQbl!YQ>W?gvAfyrEZFlr!e&!7x%2B3;tLZTM-qw8`x}IfK`_x6ig3JKwf&%W1{Wy(97`XeJ420;u6DgzhfG8T zCp=9A?lODrCsWteM4P%AxV^)F?p7Y8kzepoAVdi+Sdi{IH>P>(BL?NtKOuP};ZX5W z_eDpQOEci8DgedqB)EZn(Kxc3zUb+w;FxYVMb+H9UU_P?Jsyye4e?*Z9MWtHlsJl# zj*a}oeZzFjnFHI#GOo#!m+rj}5LjjIeA?PC*bu6VMx9gs9L+pJZ27JAC)6Yc4lEVy z+f+oHqlrn$fP@DZ-&(r;wPyi&d6Jgx@SCCkXr07)qm&Um>2w839M3Sz zVjDneTes-)QN!ZfBwt}0q|{vkc8(e|YKPL5Rr68yPk(Z#bL)aHG^4fx z##KK?AtPSt#R+G)&)Qp6vG>-p^*0tU^mm1_8I5cHy9itHQN|P|&0Ff+Zi?>zOvalU zj2-M)Y>+Hjxn)+~7qn1LkEb3MREFq`nGHXD_@Y5DJlE?Ru0_0DF}vxVF+m>PyM^0*MsC< zK_=I08REI7Q@h8A?x=gqaCgry10y&9LIJWGn;X7Sr9fIf@lKRz`+211an)2uK7LG| z_e{(V>yGn3)qA(Ss09OpSwZa%@SB#W?KiLNAmc-((He@KqPBcPS=ixau zjvf1#5C&}|sqCfv-~=~vTmj6yyiclk1##J0xQBAT-^Kgw$vyDG0G51F@6D-I&btC( z7JA)*IF{A-M_E;h)TReQ)-PPrpTSU_TdDOwVS=F$A|%pu=sAV3c%H}QW#6Cs$yz4% zRr5lF%S~JsyxOS+KJrQbasG$rGw6RKyPrC`gMq;CEcs>Bv9lZsPgr=mCmc@%qWk`^ zBKvyWIV_zf-7ELO7_P<#1|Co(=o@#q64>T2j!x^y{Pulyq-pGtLe7zg4`jJ{Td~y` z`TdU-XoQ_r-#bQo0JrYrhHjs2QI=D_JjP8O57N2swD=@}w2 zclC>97CG-7YfLa`esfu{sz0{zILgUhWpokFs6^|OP;P48nDk|dK!qvvz2Qp9|%Rj9RUl&NBRl zeS6-5SI4}1VfsRMNbb!UVlOHq_9b2Nl9Z-DRQ`}Yn0OS41LxZi%?F~PdKks0k?_*ait8lpEmvEO~ z@n6efI1%?Emsxq`eI}@*FHwt_WLV1lt3iMDI&0{4GenW`-XV9{1n5> zTzlrRj^`UpA7oS$%#(?Zw^B46d|w@Dd7Uo>G92mV+_W`A>FXqo<+BX~s#Apn_;adf1y$hpE`5=!||7AdyB!O z-ch#LPMU)ei-x)3CoA~|yOc@R?R2=9JkI-Q4+!C~y?m-cOJo@Pe)a+g30_N&5|}8BLYx%8K|C}JS8(g~(%A#A_^EL$);@#FM5V`1pGF6ljFo#l zMnd`RkGK=A4`fLtx_^%V!F``IzKUi*XUOfR=lic{-}9nm0&M`r<@}+170uC;y&!F< zYwqi5JR(RZf)p)7Y%0F~+r5Zw#?uY=cg-E#yx;ExvB)pn4}R`(56=)paPmf-tPrSo zr1FRM1NuA#ErPBiJ3v-62_m#Zbx@2gh+a9ArtOUKJmG=b+_je0IFYH>JAsFSc(~{y z-K4ZSl5EhdD3nzOSkboz1=^UIW-nc=KHv9Z{NMu_TlUmOzzb~6Bt^yp1bYzLyg2uE z7WL_{Q6;_Z3~x5Uo^?K+;N%XgV3TUoJlh{vBKGQ^y4Tje*86DXi?8uRwU>Ji*^F`K zGUR`&?r(@rr5PLCgZ5tFq1}*fXRVXrQxe zfa>26{4gUQMX6>+lFiF_>0yuhJz{OAtSFoNXC+hPv|q$r}=6`(i$8IIxg-d{y>m7S^rZUV$a{=X0J!_&;Mud zG=_Tr_>2E1c&E|4#j7Ev|9GdbXP9#N%D&AxETRxw2xPO=1eh>V4f?)tQ-G6(6|K$hX|MJ*kEIMXr?VG+9@Zrf zD^0&ami*^4$Bu;dRSEg*_GWb?zF2M~*|%CjA>w&;7DC@%0uWsbFqR6rLB6GqhE^OjcM!#p%k_)nPpdQ-6zjy0Kpy*O4s&nZmJ|hq`hiQzaLFtQ zDD49@l2NhGC&!k5P_-T%t4Yv*{}`F8E=vYczWw2ZEU7@@3`CzdsLa zou{x5?)_iy%zJo%0B-y47U!?>bl6V~mcCxTPD)4DnDw$+vLEgulZ(DQp)+JDj9g(v zyortmz@xc_2&65xa|eQlS6r&&3%NTAL?Bfb_v#!rU#K1{PoB+sdwY?qwg~dmq@Ae9 z$J1g=lHej#?2$v)(mf{bBkL%?rO&VcQ=5ioop+vdm9nWZ*7tUawk!8*Kl(+|^kF+v z@fJnxYQ0Ojg{W8TS`GPgN_YlH(uEf%9+dkqL!~_gFv9+HrT3S^nP`NSqDog;&>+Cn_}?0z(8akRg;8j`5mNN@s8yz-3g0W--oum%~% zM)90z3`Z7A7|3vBc>pY$0d$92TC?IE892739g!tiLdrrppQ3y%MRg!=?6P_1gz(c) zd)$S0^fg)C^1k=U;X4OpWcY@{#g!iA_X~k2Gn04m^CYiYW3-q}MMpDUo-k6H)9yg$ zX~^z{XBwQ2`>5;XtP)FQs9`&OgDOKI%jl>a(nphVqlsvb`h=cD)5MMx`Vk5Jw9I7) z)lFJ5WsgR8q58n)Nay0)-4vv+35RgD#YMf;ZM?Rf6$j_kx)bzY0-#-?HN9S=xY49| zjVBP!+CX^**AY)8olme->{M3MJA)OLFecdM*Dvp-w>JARU95V7Z=qls2`^ue8qnvc zoz}CdlGYfG>E3gCU(x<`%-^WFZi-^4`y=BZ=~Y|^vK1vk9R?z82=d2SJ%(Kur0^)? zqOJ4}RBIFm(BC&3&04OH>qR!B(+VBI4} zx%Qpn+k`!-W0?@{kgI5$V@f7<{BC4t*R({_uiWX z@!Uyw{HbNWu)G@mb~5VKu!lYHR4stFe+3yD(LO+F#ix^ZH^b-=#fH9nL_0NDtQ(TDqq@(uzQLTzoF8jO${B7rdXCQ`{LC>u;bzJuxo5 z*_jn2Ga&x!Svqs%B7mgQUF&DoMc;C~fBO1!y91ALKDQI4!$QH5;XVU4 zT#ISGnsC#F9fj*MnkDzK_ti#MTCx4W>HbXc$31f!5*own%}>Wvl8(8*F!j+5uu_q4 z;+wFAcPY8S#&*B znl5#FcV!YmtA~)nUJt?leQzow{p@gbOWH={0SU8Jlosv7(RT2yq>>}1tTv-_MI99$ zx{jEYXNNuel%Ml6t6f4)e7)B34i^;dbk8_E_sil*>!{ioD2R!J$!I-o->re3Qgk2iCQ@lHq3%NMGH7{2jg$ERtLZkX(`M(A=r9L&Ej z(&>=WW7{Wr-u=iz{u@p=O~hM8{u*T1q;RIE`KvaO^O2TF>~;3V9?&`UCHJ1OS|xH< z*a3=8?xR*4IO`6Xe3e6Dp6=Hk_MwbJgG_+0nywH7ZZZ!ydKuF&LXPxj52+5(T}EPc zUoPyhJbg`f9PkcVx#~qF{cXh>NSfxJ!*?@ft^kMg;W zI814U15$Sj&HR?ZRi@Xu}JsiCgfcK57ElKfy-%8%7CC#%+ZABBr{8&lNvJ zSj4w4J$1>ltcdO|sk>e&xT zR2Y0qPp6iKqlA{Xx_rg=ZrE7l&n2lj`E%NPfGZy~-Vpm96V<#KHrd2>?aUM_DE zpuy=;(7crn*|jr)aeEqq`;9)vo9MX~*mDWY*>0F3vF2$;%sgb)4U|ctVu) z0lw7+vv>bJ19*;{b`FVGv@leV?Yx7E94YcyTQNgehbd9}5;Sh@>Q$djR!I36D2Ru^ zC{k#?i7aUL;SwW`@-e`tf>;KfIm4BT{aTC1!k5QxKqvy>t8Am)K8YLOKD!CX zdK+}Xm=-ZXuk3B{6LMkU2BSz9y?{|6P#;1d73G`%zR24-JJfzywsCg*hAs=YRpp*_ z9|$qTA)FDizJ-SY2wGBNkUO^>VR!59!{|$}V#1V*$oh&Ng=T9X_C1;}WiKO<1-HDI34ExEkn@`2okO^=i*gC^Rq&09SFW-wliBA zycf9saN<6J-u#r!0JK)yS_+bS{vHlmy!xu{V6z%y~_}%gI|gVcm8-updGQuW_6SJ6=GIE!7L-BM?OE$CmkaQ{`o|evTylM?M_gq zHiztMj8BlB(|Vt9v19Mp-~QHeDNEsHzH^T8j2KFf&u$sw zi7%-@TmL~UL5+a~MdN)FJ#}p>1`!*l1J{>6%knMKRfarWm70%!^5&sBa8zaPcyx=Eh6r8~}hMjG1v`#ClH0ENf};D8o49gd3qB<)N%Sph+W z;Gm3H#%CeO+F**;#AsFN)X)~v#(#qZ?)C>LqOh_+e%~k1fV^nH4`l0pK7!Z7TSyG5 z$ifg+M6DrDr}ZmI@b^LwTG8w4GPqYy`pn>JoKZz>TZ2)Qg+|Y5A`<}}6QG1$jGF{I zc8}$ZJT8d;Ijh7#alhrfj^rx@il+*~gpWVTrzf+o)nyGsTM2S>7T=ykVh?~2p_>|? zjPGTLtw(4=Iet+)b8m;COrpc27}CLgMDe@aqG5inEh~Of&fwx{Yx`zz>)O)QMZVQs z#WCSi(G9ZGvW0OOhoO|0>JxnSEj>_96nFJ*w;!Dyr_>J3uI+6tE!14cw*#~v_K6iw zI}kAyce^z)e%mqym!7FT|6yqLfm1 zKNk`A8$jV#^n;Se0EkPblXrcj5)jGn2kmjfOL4C$u_Yy}u373kRfXNAWS|E!lNBAq z*`8JPZJI==>hX-?{*J*|i2n^>H3zJqPfF8#A;fJx1s58CDZBhZj8ah?-F;rJT^f3G zFiY(8YP?(QM!2;z#InY&a&<<@lu;=1zy{YYNSB2euiHh#v8Ex)yKaMa25(%=)bqz0VYh=3OsZ zBKO@N8HjUq*7Z5=55@P5-ZHWJuj1X$c{&J8Js40#DCRxx(UjO6i9y1hpuimO z?6cra9g$l33a{G|su%Q%05O1@9iWj<)qHQCcMa{ph7?^$*V6H$N5;9a&QH>JF(&)< zqVH`wz*MkdtDDO2<7LyFf<{IB=a}*tDud+Ar7ATw1qLmQoY{p zLyZA+c|D$^OT}p~ayWdztC~3FwITExya~u_>V)+e%Oa($qygy%8^>Z|&G*kZH$#n{ zEyVjdG)KUSh20ioJ{wTf{h|X0PhtHEG3l!APvZIXR60}|D{NWo8E3iGYgiQPK`mcg zT;2>obKY}+&Fq8m^~mJQ;uR%@yuQk>|9359*Z7$@<;_lJ*nymSw&&7_X+nwhhPc0v z)C0|IZ?KU&=>E?2V~pHvoL(ecCWKqDcuO?rrX~&GgaC;`JB+M(Y*gTr(0*kwgyfKL{R=587DqDrxNpCuFc4F4)aq)7*}JWt~PPXiD$`$&j2=9!Y|EbTsue|+Hi*E zH`1RODx0SJirISGKm>ey{A z9rSMDdZ_f=x&Qp}J|ak=AqRNvW$WaHYWZBL(nN}N1`9=e_J|=7j+jeeAMUI9xH89hEN{aOvM2g;MVbsQg!uKGS}~?*Jq> zUJPQ||7CVmELke=3W0jrBU;=LQF{yi;I3ThD z?gNQmoi$5cIPnpdOA8nYe{BH(sT(YTBL^Y3>4jVY@)+8hY={Q};EIvn1IA_rv9rt$t-?jH!*p@64#5-s*;$VT7ynITMTU zOMt(dC{)foC!|TR>zR8Gb#JR^S2Gt)lA=ueE9T2CS*Ffty@m6RllfSXYoM?T}fgIkB%?v_ESxgmVyAjpWL=BXX>hU~ZQ1cWKN+lSZinan5BBUi}e z3uaUK#Ve4W{;1bafO2qEG-B8d)dnjSi2j~x2dITkp>Is2&^Rww+n5o-_>9{y zs_!*FTOA6haF5y)Px-!D!gzGpLg(;q;D2-D6gjGYmt1eKKDS@^`kUbJ*_z!;0BZK} zI98Xg&9H8zPCKsGCu_KnKU1n?R}a0@O!Ls+!_+I;TcfKQeVMJf!qecTI%kQ$ST=tv zfC4spH1I7(6;W=7>2xJLB05M?hbH2;{f$IKF40PjtzID3t8C+x*8F(C$x~DziO{LO zk4YdsSK}r|%}yGEg*q_G_uI$D5sJ;Sf!a|$6@%xpglLcwSrxUCAvEfuMNwqhWf#=k z#bz*}@i7t63bZT>1v7}-@x#Qp3qXws7Hsp}L}_&``W+nLAc|qqn|LI{C0D@1|yyneOQl`Dl(PIAt7Fi-zxz`K&cIIau|*r!Ej# zL%5!mXv(|$H*!uvZ&LFHhjq$_!MB=Hx^@9=Zh?gkA=TgOLQ4QUcK(AZUL|^F%Fle+ zh4#gIa7Fjy#RHYmlg-bVdZFI$8IWcFI&v8@WaV)Sf8F*|@wu-Xe!qh})YFf`Uzl#( z7Q4;H$9@~@dRe+1#W23U$@D9JNr|fJy@h?+7u3>i!_hXaaRZKCIIW`zT!=i&wMpGS z`vchafb*Qp;|oMPMEZK1HS+dW6H)B|+zf+{UB41Io37UIi(3Z*^6l)reffb;q!bnb z>2?{Rwa>%ehaB$Ea5#2LoQ$uYe=*ZN7WCb!Z!-09uP+)?8{u`(UVixW43kxFH}>W8 z`EiOI6G`!p9(9yhygQofj^23poBI988mA10)G*&Z`Uwa!oxW;0RemrhG^_V0Ir~WV z#+YvF@zyIApWny*41HUCS1&CwVvK*O3<9H4Q}zoKqkH2M>7bRghOSclkfu;DC)n7E zBr>d6bjc`>84Q!2$0Sc8MZPcx6IKC%IfZnJlEpKx8TO$n=fu=JxE^3BW2e8ey$2#X zG}YjJ+?oIXagYqi#Yxe+Bkkrb+TQazal}dYBxDS|KOx29>P_~ zT|op(R-nK2@5Glp9``txf6E~MTSIqtsvZ9QPv*|yKfyJ>BAed(ks;}aMq#i&=AC1} zzX!?upI?l}e|_QgUwJ_`Tui2#{ISb(Ty7a=tjkVGIvCzF)0@q7IC3LjXT-@=tixduss3*TGHE- z8l0;&xnGbG!oV;yOiPfekwX6YLE@wquJ+OUCIGNXKKJ2zFQBKO+rQ)85fPA60#)cl z@Rgkxg`eq_Hy&OK`*$8WVOIqr-JNEYJJ5Ij>s$2W8!O9q3@4j=b3Krdco6O$p5>Gj zD^Twe44}BiSN0AA88d}w=yy)D65Wb%9F^SJ95^84Q{6m7za*2*My1C#xBaWuhy$UL zJ5slcZKGNa(=ISjT&fe|9B_iEX1y{1YTWP}7`rW;7ts@{NXjG=gg}|({d?hEu3t<+ zZ+}R}D<-~uyMnm4H_(NP9O$odNBdqh0DI(#2f*dj=Jkzv^glEmCfU5Bbq%L00w#$O z?&jBvLw>Oij1sy5=$<^>qxX7SoxNSqJYupNv$N|_d)T7A38Qbz?psEV3=k?GAgYKi z{gL9heQ(|K)}IIohkp7{_^PRTr@7n}=J9;nkx7G*sY1EncLqPGxby0M0mx2@4KvQsc)gD?=! zV;M)wsn04`XM&(SfIZP2g1k>9FtLv+MlljMfbCu}{O)$kD8wCCi|Obd@H1J^W)^~{ z-RK3E3)dC_jDP#Kl~a>PKVv8b-fM)MmaF#;I<8hcWC2#Nn3*|rp!DJBM`bAnH9&du zne^O!ueU|4cL2K8RI3IMQq}bULV%}0ozDRt@P~n?3_F#VSqA$7gFskn>!`5gj!vM*IkcWaw z@4NSQ>npe!=K<5eF}epU=D}L?fn+kQsF%m4h(aLj-|_gm3oZd0@B3kMDPpU$!`}a@ z`Ra@_k)7)qGEQ!M6y&DhaDm3f1NDfPA_%;k@!IU5h73K#-&OYx* zopLnVujI7p-ji}h;%OkG@sr-t<5T5P^^fZ#!5ke^*a^)zB>^5F-r7>F56cqX@KL_| z=>9~C_WTuJ4-aQJ>At%Mpct-YmI*M&`&zF=)LHJ-ZOvGKTOo9{zn&11hin3uDo6Z{zJJ@3cCUFiHqTjBP7sH?-f z*7#<*c&Hzt_Nv#^Pm1!!^-r-|q7zb<(DErlsHBJZb5CFMeWB1T=+?r9@>#*C74B)D zY1URX`7pec5>!OPrtHH~d`c#q?8syN900!-hbT5*3J z!^v7x1espiJ05|en4fN^F8OlWV-SI=GS7}n8^NQ)P~%XMcgRftSdz}CQrvuhByd?3_p{~fYX$5w`{#$<=B)>;;G+P_Y=6v_K=#(EFy{k0fq%w9WH+~&KedPKULigYw0`*uJ*bpYm4s&% zb@!E%-fhOi)eRt&eSosNVWpx1fxc$d&+&H=ZCB-1}#C_rv;R^+{}TI=?Wd0S`nEcvcmtKGkcr}}t-f6s&?aY4lK z`C)zUO3gFOS>ejXC7}~|J%fI!?lN%SC>>ON-}(K{Zy)-B@_A;TL08wfUdSG_5T0XI zuCMl6rv7<(@OR=^t!RD5csT!f^o*h$W>~J5wjDAt4%BXj zGxvGL(g5Xf2B8L3MND6+Su3qy6$~_qf}sFc^io}b%sRk?(2Nk4EeZvp9uipHPGk*{xo!sgTGC(J->LS-Rx>_kG=E6wq#5w^>lIR1| z6(SWqBheITP~JV@Huw8fzLbOFalUSZeBLW`iSdy13wIulbJOOdGm%}{kBfSeC z0t`0jb6uj~0Oy3$;)0Q2a$XL;6IdBlg$o9C^k=sH2hv{W-!gz7}w6jI#rH+bJz1DE7Dt2b}G{FIwK99(jip#%f%5C~O zo-b!a5#UzG19Eyvq?u?hC+Ph4raU76(%27T+EQni10~X1Wr-&Z@qKL$FUDGX^DCY! zp5bp`zC(ibY40en-qHLy%SYmeR;+~i_nG%zi0IqVUT~QB&$D!YVnCsz9gQ=74MGGA z>`4V;@W~^@L*5&iW1C8cMy8vtP+TIl;0G7Zr#Pd-Jy_rJA1?Gx#M55&D zm$t*wCN9kT7io*1pQ7>hfqUcWuzmDm*XlAuxR|o_So{>_mFj46EP+3>uTTF(Uo1JM zMY|8Mw-4Rpj%JsybN_UP(oGCeo{i~dN*bcQH4y@*3mLHb^rJ3DYb1FeD%0TEHxe4k z1C@JKRdE2Bl&-Id#i3_0frIoz)Zr`e!1;3O0@P z8zdrB7+~cxrDx5)z%)998rGv+4Oa}|aUjg_pxD(Vo}pd@5`B49z2 zzjuiVe~M#jU%KW1X2fgYM-T_oT2BA1Uf1*ILmQ3;a@0>|Q1Xg;4Ozl+IsyG52AF>w z7DP^e9YdH|Ssryv_5&#avKkMaBp_Kxb7U0*8=R>8>7Yp>D(VdaTp6+9GS?SPC zlotAmQ-(dGGhX&Oe{Z?XTWO)JaV45=F;?3&Gv@`di#%d_w1K;8x62qqnW?Cc1CINn zc0qF!KrY9V1|#7{rbq9U32C0V)ofKz({`!VxCuN~a)hRVJY~UVJ?t7Q;=0?=6~het zyanmKD`FHppypRH-3CwjzOPeLL^r^}3^)qh$W!i^y#s^NJoDSe3umw6+Xxl)?n~~A z^uE-Z$A>hSu8}XcNw1Ef$phb`LntlUe1h5eO}@T3lLN!xp*x_s!hTX}_Rt3OGwUhI z!!j&neP^$Nq#=0m{wgDox1@P4RkcMZ3a*l8ra8<(PDm604FbFo27^TIS&bST@AG51 zM)W4{_U9tpt^i2^Ekzlil9#>mz+O%-@JSKM(NB1yze$Nk!~I_JL+z7hR^SZ$`T#zO zyA?=1L*#;>$2f04;wN&BAISrMr|p9}`qvq~6 zH`ylB$VogFOd^fH-=ZYegs888S$@d4fLAVh(<@qmsk1MLQ=JEu=rEws*alkf(fe+M)TMzt#{FHzc8h`sW z!p;gbaX$A;wAp4Ky*}IA0ipN2J1XCN&)eV+=NDJldwV!OmGG&yd2G|&o|N;)v%7pA zmm?}ad_V6KoKz8)9|rZr04) zWzVAzX==EgD}(q@TFzG=f6NR}bn$Ih1GSMu@3f!gV;)Dz1Qp@#p!S53(0(KO&$v`X ztUc31xQJcFb?G-oAYr9uG69W{d)aLVW#|Y=z=_6Omd9$ z=qMlkV=loQ$<{8rkSysd$NXRgb2Ka6O-oj|Y-LF7 zTsMbN+1*OT$Um%T`RC&&J!*l55#&5YrKnQQ9djMw2}ZkKFHmF3Ynk08IXvc5be`9h z@qoEG^rX<2d{(veQuZB#nf`O6D%CYK%X2>;Gsnh%-{?bis^HVP0sKbYf9nf_jZ!`K zxmI9v`goPXrDHDhgMu1Va&m24Ts(S@0TS_J{6VIN3#39ks%%~_JE}XS6j2HS&uQ*U z+|vB{iqJ30j>9?+&jm}y$#+FUlIhVvve7In3yetXD-^4U3 z9_~c)*M10h8WKMlQJ4366{-}(u1@{5wDqOCcl8$%3|R3+T{J8RmJGeKbB@5UifHfU zpGl;Me>w_z!u&%&2-oX@JI2u|jaKWyX`7k=%lBz|=8N#0!M;bg*ofEyq-u*I(Wj#@4b5~0U^4(4qHfiIyb`WBO8``)32XsYVy0gg_deV zGK3%45VRNN#eAo{8~(aKcU{gc=_Z<9j#OUzh7I%-+%i`WcQRb?9fP2%BYD3}3gc5M zHeLNQ0>OD5dh=+O;v)4)onD3NfQs}GguJ{6h~&-JZ6KrE~8GQ(!hdu+Ci+p87Yz#2mO% zuQvD86CEdTyZcfhI*8$>5ghs@XrFRuUP+xOjCaDR_l}xG(xu6~D9;-V;akmi9a1e- z4E?U39&_nOzUvEg@1XbZ=h_=I%i-^yU*=5z+NR@`_%=e1SqpDNS8xA-mA!wpvxR4O z%Jf!`TXfzR=tl=#^_AfV%UE6wv|#BKR*pWxx*1<SanC3Dcn>>Zl8B{vAL#W)daFi%QLuHgpeBiyTQ1s2JuC*Zy5D znzf$N8ZZCsJdWh$n~LA*UH+aYK&dAoa>|HDl`4MCx6r50Pw>?{w!V*q?`phaS#S(| zBf=V|qZY(V3%ESSPYXh&U{K(XY<`JLvu?!3ThA4kX#{Y*pJ#rUY~@l_NvQq= zj+VcE>z;K8M|??B5587C&U>NQwdJnG@O7>%`qbMT0aK#$^t62Q`%^T}T56`MJB0ru;$_1Sx=tpP7& z*fo^k^)JUS9!t06QH$0!x?LZ~EZzl1YlURN11lTo0ZOgGifE-O$QtaV64Go89$>m7 z3@g|s=0GEICi~&yf=esW%~!zG5#vX5w$|?NH(a_x`Td0w+703j-vBd5kp%VNO1o3= ze)QM!p-I17pB6yiL}{RH&lKI*v(F}+*#qoFI3Omv6ZMLoOj;tXZM=6TK|f+Brr#LV zTlGJfDFz0=(m-cm38e!> zkKZg*%9q1)g9Zs1pE4}o+?o)@meBK!G7DQd#jA2Hj;mk@dC3sDsQ_)YWIRY|?^7f4|Zv#m(n~W;N>Zlu!P8 zT;0NqiZhkUWrNJY2T?xTpd*6-aI40oPrFJ1@ZtT|Fjk({t~_yA*XVdA`yY-wOm;n} zXzGBNXpvUk=7C{Q<+eFZbnqS#mm~!s^gArDTq2Bn zyq*u%mRaBD9kG6_4}kks%$bD**jP=1c2XR0A!iDrx;+rKvul%@3_f6If9cR z{?tco0Mq9HRyK%+T0Fsmh^JqGBYLkmp&mW+-Mwi!2REOD>cMK0#Y}IC-{L}|xsAtE-gmiY(f}IdSKnF5am}og@(+ZMVEpFU`??omb0i9)3IcFF zZQ4(_IqAML0!1L7p$+4FZE>LL7YY6h4P|z}o+BmKBkMpuUU9Cdf_#;y@v*wEfdes_ zE0pOztOCrn&@f0+5rjB$%v`3fQ=`Eq9% z&&PYfkdAt2883iL@YuuUu1G0zOY3l)K=Ab7r@mUsEo~-?HHkMEJz&k zxrxZXQ23~0UQj$9o*&l7pvu-y1-c*T>R+pG-5z$AUTHq(qJy1IW3+3u!Zk04`2Df+ zHRxE)uIAD4$oI$+8*GG`pr_(bLuzaoYQyGB$R=`%P%_#hLuMj6VaI{7m$) z;(z=b&x??#qKdOlG|>X!`rXk=_AtRXK9u;|M)q?Cduzj>v4h{1)`xwQLI(8fy}Ohc z)If^D^sTBd8bq)Hw47ZNb-&Wp&NdOWXc?fbF8?S*SSla3vnwp(u#`nK?oPhZ{qz%O z7r1+gp9$jU_s{z7>uCX@p}??nSGw%yd^WUOoMG(jveI2`zRF$Y#)`O8gG#E{Mj;t7 z6z|K@XyK-`z=I<49QLl~xR;ONyu30KH|(pAB+>b?n_Ylgp#F5$gNCVEEVlZ^D36Hx z3Pvh)B3WUJa=PaboBqp(xa zIL?U>@Hr;x_2q+hf|xL@SFWT4-ebc~PQ*rl?wuozL=Nr8oOuApO;t#7@8gTtez)f9 z7zESKU_oDfv0_X9OcW9rF-zdd;0b;_0nZupu1=&*%AsA`Ujf^#PdmTH{Qg{oI&`%e?}>w(<`3Ql2Q52O8ji?iT7SEj#AunFgH8zq83*_85-62>4+JFDdlsr0<#jx`g+o zkKPO(O{?FL2iiTb#Y2myM{>dFuy5pKwRn1S@8Vv*k1UR~*HY&eeg-LNtlh;}4;$)R zr%5Q?tvl(N!u&hHNtttwcz!qlpp$N6e^A1R^xZTuI5Qf}V*#`}ZKk;VEliNO>eZ{i z?B{WHcxYG?T5I!f7ouF=sonx_+%P0~d-Rp{(A1xxlo~oQbjQa7sLh;%_ z+e^6m_x2J6kcKhTvKJ)^=qm?Nf|ix`-aP0+TxYWkBh$2iiS*N;$4_v8dT~>A+Miod zk7&obB+nZV;&c3D9ybPS@rovYP>yzX)vKX7zl-4UVZU*+IL+m7*0R^W5CPFqrS}>) zYV>hL$WugG3ZGX*MUZOOJGiH-Gs%`=`FcI`>-M$qw(x$CXSfb(AxaR4gh4oL2!06? zv;MI*RtEuQ5aZ*qGTpqB9+X}<0{iKX5;V4S zH^0KZsvHNodhOT3OFc$?6GUlcVS@3s1m?qzXah!5DLLH*Z5A;@VLv!}+jchyYQ!Qu ztmmgoAO*&&*CERI7cfU}_Z|4J+^oGDq^<=3yDGB9hy0Rc6*jiqev*kIqDp)rM^&i=CE z2*nWA+LCs$KZCDI8PWo5u>j($jVks^dd&CE!SS%~$&MF|deD&_7TTbplG#i@yd^j6 z%=mqC{qa#`pH1X662n>*?O69Jva_c&y!Ad2V`cx>o)_RgY*#q$yS0ukb<@lTsYgsJ z{dl9hxZi~EFU^qGv6_+XL^H%ORKa5qE)L@E{B7+?SzA?4;wmX>j{EqR!Om|?&=ixR zpM|DrQyNopnql&Bns6RL$XSg(uQSp@yLT`?A=<_H2>#+^`i_N6Ul}jCqWr}jUX;Zs zIM7ezVB?FV5;I>nY4YP^GZ&2;i!Px~T6y_kR}GTL`;EX%0~##y2{h+v^fDz20J_8&Xrc+~uaB`EywRmJdjQc{%P#eV8Q7os!Orx@c8mOCNX2N~Zv zS3bl$GP=mUjK`Kd3x)dzFATO8mciVp-3!7u zSUVT|_m89=T@ib`Cx>nXav*7pdSSt<2+M2buVT_} zkmhzPwg{5~aYL$=dQOlAyn(iU=bH%`E~Zhaya)~z9G;vi-5z3`4cF~J%jdRqYW2@t ziLiI{&ytUbGnIgZ^&`tEb`o&MAXB7D)7IzLiq#Wn(_Z+rqg9V3$PWoh=Orx6d>nXD%}0o;~6&HpVK zc+7+>HeoSh>K#c4Xs(X~&H?BL9?Y#_T=F{IYC13+T&fLH-5!* z&oXs3<9ge>L_uDH3+W=z&ymx&DTG$>jI9UNu3gJEi1X85k;I}^UK!*KBckhJI8VcA zKcm$3na?FbDWSX_4oheWOmQ#W9mht%bHBL=%f7G+nLmSx3lS@uPZ?^SS`X2fFB+m}AY-_~ zEGa7k;#Y_{Sb)+)F!N_Sah-y!3Z~DoiJNZkJiVavzW<9EDoKlU>zoe3czj}(@u5pL zude&C;_`c}FuU6owwA~)&Yo?t?}C5Bezl=1i~s(3T!ZKHg=~(H&)nJn{T(P|+w&-LX*~${rb`M?joZDXH z&nimqGf&Pgw%~pHZHG8kwrA)=DD*T?o0&`Uoc-~k1-HzQ!;!^ZQZ0zt=2l9n4?uAb z`x5*Dcr5rm^<+Rjj}9D*#gq>M&W6MW8E3d!lTA(@>?QG*ICk#uYhr#?`H{bdHAJf8 ziDG~gAGfzBKWQVwp1-%UY8gJ~^~{6pZ5EdPG{WXC8vxhR4Xuzv=-q7P(MnAuiZlA2 zcl~8>5O^a{%|nv{tb?E~OZ3{K>fMOzYd^9;JZ#+4nP4-WHR?Ta!)s0S^_4Y4>@qDB zT`s0po9^G@yLssV;kzrBsp^Fy*SBH ztg;ufa#wG8Re%l1pdCvxP#)*J>4snPSFI*6kR$9y0l!P`&3*$Od5x9*vZ zksZ4Iy#Jn|d)bq=#Ch_^_yMm(9S01<_9iUI)c5r|L4Jp9oay<-8_^v>gb;*Hl()7VKgf+-eD~J?~|wN3k+ ze0W}!*1sR}gq2)47E^O^wZpc?O(zG;MjjpbANF@^u*CaTtpbm>+qZ#(qQBTKlW`k|fMtI#FR-|elL>IMQ3*ZJj8cQ~qej(s*GYjf?Pacp~Wk*)?L;+KRMhOtS zgjd=i`G_~e8v#RJhX>HB{TkD?HlAr>P`_+yT!V1($hu94E;N(<^UR`3|M;M({E^%4 zw|YQ}J2usWD(kPJk}rBbL(`z0_fPdK^iA{{-yzl{)5`;*=%-r~$F%tXxY|m8%qJ2w zpcMd!hs1s|COE^9fyvwUI_LBhy0)vnGAkcHA=vpVOP>z8Mjz;iF!mtToDq|5YK*}DCKgv-8P~lpX ziqb>qX9EUCnO~YJL-qAlsA>c_y*_VynFh-N|FpyS;#}+y&1NKOL)uS`1PCi9d7ZX$ zyiXUuFZ*M46nhoF({|iTZ#?wX%!Pv?ck1tAPUUg_ylHPg)sLA9Z6I-m(nFmL`D4?X z{QjF0|Mnf-eyMRefeRSq6b$woC31w|cG$t(;#sKxGWq+wAl1hg{;pIcDhW3U>EmWB z*f(M>&G~ji7-;WYltoZk|E6Y&L}4JSUl4l#%h#d{?XMkvl|!<;pCU`IF#POFswvG+ ztz%HwcJe;zrhqxTlMJ$Cyw*6aF*Bc-sW4wMVylVbc_+xWz_bCy9@vk0C#K7x@RZ-i z9rncud4fSwW#&-f9)CJtwf6vEz)jz4!?*8!i_35Va{dmU87ZP4io2zo2~gheS`cZo zUA{IiLYVj8&)vpwi?51qaVFKlmh>ww+BBTEjV(C>H-wW3oI>6od*oMkdKzxPc+VcO z{!R=)Mrn0wzNy6r4H6PsHdOg_U+A1t2tOwM*vX+{ws#}J8)4}x8`q!Rea;v`@h^d6P>jq!67&euX%xZ0G2j}62HkcdhW}fQ(lwo4C-NPh z?#*^2_qP^MdlMvW-#pHKMMk6&n)Z%qM)4yqw+5hvN<uCq;@L`9p zIP||-QNdkQEu@uo(?_BF0Xy0Dc-pI{DIasm81qkf%ZP_?dWYv+9IjAwyYeSKTZ@r_ zwQ*{GUr`3-l8bxxC^vu=dF3=E;?R4eL4BKGzm@;?#)5goe>$Lfle@-YEe!DSoFX2~ z2ZqV0&H1tWL&?$F1X{+qiARgK##S~aQr-DG**8#29C?vjsp=q!WfxO~qKzM}gt4dB ziq{PPi8>^Hz1~l!IcEKR%EB{P%y=q|e%$yg@(sEDPJILses=5c=oT@O&^?keR_)%l zJVeCL=M!1Gd(k=&`zH8=Bpq*XmmXf=8@aOhdBi}tXJA49_Q=9r2qWj?b{bJK(;nM$ z@7B2tEPUeoOI@_R7N2-%MIaxKAZ+>E6aE{}xYBX$-tE17x_)H9pOLd?LQGU0R#Zve z^^vGdzY@B8TZ;ZSsSWaMK8<)#n%JIS8?@lv!sH&t(z7$>``V->Wi%lBQD`Zbg;|HW z7i;=KrIAp(9z#S*tf8CEc0KteUDyNe4h(OFO;6dJ;g*roQ@3B2AZiUy_OKT&UNZka z96;ll` zk~^@M`P@{{uPYz-LgKs-$DlkW3@hvJFmBslT4PS@mjpuc7@M17dea|5(;H~R+ z6<_zgA(X36Ib)1$QUWZ?B%oZp8_v25%!kaSD zL~@{^_5yWlEInt7<^F8$g|PKk^qvjgNZxoc-bGRW_66VbVxd?6RsHdFb48kSvtTt*^h`^rD-!mmmr;!lZ{$?X4b1ehG_baXFB; z<$d@h?r@)`P))g7yo$8t{wDFpF}FwzgYUJUHx)6&cu|oW&IYSKZSl;Gw5y020=G{l zRc~pS(8mNeKA_>tg*%ns0PNH^MShZIUz-pgcI?wi%>LeJ4NO27?7&8fG@iB|b`$@Y`fBe`v_-^zmM3zefaTH}a!0JZfw{=&@|^!d=*hn%#m{j0(es>naFE}ud-{A>V_!^umlq6yq`VHJ@|;O}{fIZe z(I2=QKgRI%Fa{6_rfv!M-%0qo%0mE5Uu*Rfq{;;h_ahA;Ac&Ex_JaYNy|A-U9Gqmp zuaPh+;)#E#0QtN*=p96wqrXAsMqrw8H~rEBYt`@}CLF1Hi2##Y#DE9%lr&lu0V=fqSo<$<)(9tIF8dBQjx+Lsa9ZCs!`0H0BLm{DNxQHI zxwCZ>j<8xJ=kq$A7IQVybw77Lq!FMHUSoJDU%Hk+VURpmIGyh7cC)nSJw4~l@c>1< z%_As|z!+kxbM%7t+pw9{=Ms$O2Mdqtz6lzhW!RAuRI$S$z=zH+)4hF{@V@atHCV@&$4un_XHoCn>zKz4|*H`tU2UIbEEp zDX_9oE3R|!`wjsxy2S~F1h58ubs8a5T}F2j?MuO-Y1lTb@=U|7O5gVB8@Gqy^EiO) zM<|^%HL}wfm*WS!(M~f4dEXg5=!8Ff4Ih%_y=UBoX>6Yw9T6~0DCpc*f$_;CGkmMX zy*3RHX~c|aNe=~_RAOixg|EUQblpkbUQ{6A#3v!4OcYdLKlkXbc!>`b_U{X`u_Ps| zOUR1@{98H<+i0YOELB=jF=))8hQuv-dJI(~glW0#t?*A^b+^VmRY^7Ot0NXd90_l- z%3ZRySvpgDrIzmZ7bL^o9_Oy-1$3Kv?nEAek3+h282bI2rarr26E(iYBkcT3D!)NkQ$=d`oe zn#-w(I2a>FMD(wtEB!xq0{%K8=~}L3%E_l&`Xy9v^M2xP&Ml)1r6pAF{hs%{^zEBs zME=A0mD^sxMK<8<6DK&`Aaa>8SJBbtL^XG>2_FG}fqf%%$;8I!Xfg6niMP4x@x^qq z_6~$`MLL#i1WB*OH6Um{kCD5gffihG=(mN;n2(RT{z|A*ys%B%!-WwStQfp-zrDEe z3K`~#sl7H4xIS5rnrjf3&hVlA2&(JU=bT)>^> z4QkaErD^;UfKd6z797y~IDsPEWt5U>Hl3e7thGW<{P#@}+V)nB{RvsT2c#eWmPL0x z1Yu#jwk$}|wv{KJCt4|0HY0<_c6@2jM1G%UbOS^TVmB&_@hvHwGZE zpM3@o?p6{(dhmQ255c(_;)wo}T}qjH;+_V>*#Ozsan@%@{S#F(f-SK;de z6!d;g{IttXJ&&D6{Q&)HU8kpd3zSZqLon*2lx1F#dLIptM1HEIKU({Xa0%;2@3V83 z#eX>)V~{_0WU#~tfY~s{eVirVeNY@Oud+d1&K84v_zSI)BUtFCvUh&ge=?75a_4mo zyprZSGz!A8fsw^n{Z&2RQdpb3D!$0U1#1(q z#0Rv>83uw3AW`^NK5*8ja&>*jb;!H;4Fl_bs`P|3TEq|%L4jD`KR@rsNgZKee{))+ zt^5r)g4CzC9IKZS*~u`W7nLn@{hZEGBMT%==^w~VdGw?=6_!Y?><-9wDdX(zw+R7_ z6WrWuRHkUNF7D;QI@)~O$8Qi)FM&8HQv^L(-FNE<7jXXVuM{Kue22jMc<#x^S@{re zj`}V+|3eY$RvoXfDVICqeuhr~_#$>>g)Uk2XJ>sq+5RXvhw*ED`F$UdN8x@BXR@`z z=YTu>bDNQroA&r@!z{Y5`&#d&?)CirG61Qzq~3st4P>vfZgIA+`?yo{Do!ZKLv4Kz zp5S`AFEQ+2;Qx0g?!vrk zJihB77IltU?QoHePl?W6?fKT~+(K?C3Pdbwguq}rxvByuoio|at9?mjsuOi>EIBXi zrK3@?g@XK<(Pu+dVbNRF352?%!Pa={y+l_p16SGVwcl)?^sc4955x+$&hH~xWk>3e zH3GK{Pcsf-8>+O9pjL!GGx3!R*RQAzroP6ydj%!J71|0_gf`JPlxBGBN4b-CUHf7u zXCn=z5OkM)TZ({$4CWYzy(vMqtWWam?#_7agph%gVE_Q7n;i12--DUMH3}+J5bfBF z5tvV#^+{#TH11oYnMk4Fz)K?E-^Z5-3mAI}pd$kD8cq~llIq|H-!taev6#@tKZ_Rh zbzT@DzPWI4V-fa`NW5LsCfU~n>J<>_2(|rdlBV)URd*prqkO4) zu%64lQSi6^8}ieYzs>CaT{1LXgyCkD*Q3Iq5|f@^fv%f@bVy-+gDOKwYH%^x1~_P8 zk*oW(iuCYMCNEXdNzkQ*4W>snd(Tcmyr3f4T zVY=J=j0aX5Jd+Uj;Nvgz2v5nL7P9QG_)$?!9RCv4!Xdx(P9v2^ zlz0~K$?i$(eDpsv!Gbk>pzc#z7#`muk>4JrL=T9XDcH7{Q&RPeg>GR8ywBBUQ+-ZK>xGB%xu6yLo1w*-17AK9z4vzotK+m!M4Bpc zH%1*Swl#ha_ffH~c6+g0GvAjQK(p$a8;^Ao{{mzv=Pv2?J$kg}0Q?0ttkFZ`6)a*V3nS|F?0ft4 z0hQ$gLGz8+))NS7tcEjwGA?lo&)pX#RD3jj1_Qm5z1O$nO@8?)xe?|RizMJ|C^}#P0s*F!a>D7sVSq0e&%Bc4N&yQ zjV_iKhxV$zKx_VBc zw=qEozpbBQDb{+FEtHhi>tB+?i;xR}qWfhk*G~lumBWepuqO=fGqQ6qqt?Cf>C78# z1zm*21gPjdW}3iALN)qVoX37i_~7W~ zC1GNe_0Hu9dp;YD5M?-0M(H+>hFz$2<@N*g)(QR{a6naV=bu6S>ZU-32aVw3oj=Q8 z`)Y@W-<@)ai5^KgMUw6bEDc>#)AEB$fauaHis$G=qJ}-@#m#p$WTS#GmWszKY6b2; zYoN9d$eS1JDF?m0as7{COUT%f}iS3<(nW z?Lm2YaZNjfHWS4Adk{Rn}R9I4|S+#9?_giSKOXA?FZ` z{&DxS=Z&kmrK<+lA-cdzTF~0eDSYE}pL=>d8kylI1=|(t_s@H{m0nED1sJ4ggGo*H zDG^T;b}j7eChJ(a6(_>dAdNHB^-cr zAeiLG)=#vCilkyk94(r2de>-lJE5;^t{i3x+0Gy1-LuMk4S{n7lr@o?@`%?@b<~! zz%+_d$~G6};5G8TH@NhQ4X@au+(*S2vn4F~u5k3g2Jf}~U~XN*`aO$a^Hr{JW3`zf zUyxw8&H>1JEvJ6!&aETnIfw}m?kDDaY^JDb9(JM<@T<@s@6d;BJ=2q4q@tK-2zNv3 z?rEs7@0(h+##bm2wFi%YW17jI!yQ*In4-bB+ih~YEGr|&0vbB-G;m!|(MNnEa5Fj* zZr_e7=>7|fKICF~jPetK?ACGMo%Kk^1w0iPyy}w{PjnniKziJc_v|Z1OwD|Z41QUF zXWpnAT*|wxs|#SXR2D|kS>4^r7cfQdIKDna?q?Kl!w*Ds-=W_poP2}6=NhS-3APV+ zI$c^g=e-Wu*zW7+!FnmPZ3rI-c0+WBnW>%9SO-DvfmiD|UN9bD&#vliqrSOtyqtKf z*3x@SFT?Q&tk7R~Bpz|@$~U|G5k>FV+-*@QpJ1ueH`T4j`&|5*wuZGkvC-4+qRBMD zNcr4~QwWlr`8$0`IoH41#M{j}NtVP>wK?CA| z!d3P;Rhpw27Nvbq@D!ym;6i99t^+(0put%zN*DK`sLHb{Cu8zUh8Hh( zEcZ@Dm*5=&@$ZQk4J7ABGGw9T%Gmz9hmzrU_OS7NEgT(DpD(#yx$gR8EldBdnb~Jd z&cckp|CUp^WQ?+kZHA77W)6U=fQ2G zBZBfa2h{SO&cdwkDH@*nOJD0N(#9w?Bl)vwRC10wY>6qVq5>6l2c&Yu@)v-qj z1BHtKBVIpV0o!dK9^tx9!zn}AqW3^Q#e03q*H?oT)Zx>7h!3Z~=oEVN;;lW>+lk?{ zAE+~-XTdY!jf|Q^%6)csU_F3-P<`Q#$=uz2F!ca1V*#Nh72jnW1DHV+-+G&OYQ=%x%l$4$cwZ3mk6A`A60U$$cu#kTc_R z=V0D8%|Nyjmf{FEa4(;PC(H;_2!M(8EA%Si2<#B9D36Ncp;CVcgOTy!xHtKfl^6Fq zB#*JR_POs+13s$&B`tbq0eQgeMxWlQ_kAjoCX7k`2@PNl-QRG-;zY>Vdu~H&b}njg z<;fQowLLx#X8TNZ!UTuPNt)*eaw(>A5BJ35Z4=uxxcxk)w9|gFw-ELv;*G6hjE~q zlIp9@x1Jn~UL7q#>s03$VY>O~v^gWFi|uIh(D_*$;N7qo-S1&HQF8qD7E_lE3{MU} z@O5>7==77rq1J99eIW^D%qB^XiNwpF)k%F}4q41H9mKl6B~=Wi1CqV4S#d~s$+OHC zcBIz_$6fiBr_Av1KuI$^XO||3T*m9iWoc6fnr$$ber3G+2N%^ISkl=)q*=XApo;$3 zY+Pf|{#<7Ksd`T-33=?k<9llB*iU^;ZKeFNB(%;IW8}B6d-XlwwgHhqc1;o=RMmff zy?!=i^MXiU0-AQJ_eE*V6!33ZZF&V45x$DtgH4|@)f_?yzghp_r=s=2*GP@pbZ0na zWMi&&9V%@u9}+ymc56R?yq^6n=xI18dQsjaNF>N%iml>vb*UdV2iVHsy~d}xD)ESb zvMN#J`#hJDr>3>vH+WnhLv`JU3NB4(aoV{2IKQOsX#e>xkFp-wM}1D00up#wU1B~@ zHdwQRRz>uL4gjsQA@=j^n<>qEi0Czz^M`@O*A&H@xy!A+TENR#8qz3+9a*FID}Hst30EQto=VN97zTSNU&Ulfkkf^OU8=cuphgJeJ5`H{+>aI*3T$t6kp2FBty^FKN)uQz7w;wgQ85dz3s zm-b7#hmZ`?YgjQt^kDhOfLup06dZWcW$5IF-K1L|ZuA{?fj&zz#!L+PS%xhwUdS7< z*C1z!ALaCCr*o^8Zv^AgQ9`Cj1d;7qI2Lq!B8@GoT<)utoYLxLT;ug;Y{Of6<(>XmzQOZ>7fJixY|lYdSOHrG zPnZ9==X(e`SIJ(KTY3-Mp?!aoPkSyw*k7`=c;^D`SNqQ=n@*AJkha5%w*YW$Wrz_!cWvcT%IQ}4vm6^TjT!b<6 zWq#%C;2;0x(0DeQFZQi2i`U_-ddRj)1~EGx+G~*Ezb?f7EZBvkeH=gRpA$)D{ISyb z<2cbPoZ9eH6#VCT|8wv?RsVNz91Ry`U;1$rg4kd7KaW!Gk6rw}#eVjGynDL7tM%{f z|K}V0_p?9e0#0#1c*9SB-^jk{LOyX%&N4Xkev@O# z|MP2AQ5D>e(O~J^H6K-iVTZcNf4qb6-`=6!zea#6{?~Vsj_=Pu=l-uR^i z58P5!dAzSFT91ASOlIYwy&s-izruS)0d)En^BNe%(1l3VDMpk0fKxqF}-8 zHS&O@Q>U^z`0N@50*BrlGggn=t0~*C|D*arFLk{y$rUNH%wF%l5MRCweeU>#x@Ffd zXl8Nx_YKb#(zr8z0;r^w z7g{#}kUXNT`nk$u{X!Q360Is>3&{IFqu^&nNfvmhrd8{G4JwM@t18fo^V{6VXP`(= ze}ko{57*zCA`%|Lr(t-Mw{H7dtm14@Yu`MB!qn)+krx9Q_QrJ6rX#c%%-K;6F52h# zeZIBEPet(%DdZ*k(KR?^$KC1Tm70AJxo0bA?|}d|{Gx51=y`LSg!c+wxjyW%(@|&E z?%iK?fjq;#T<)mv%9-C3Em$qw*^9_)Ce)vB#_4lQMF}^l6wXLI13@bD!fo_qpvQ*z zc>Xvd`?7052o@5Z^~6LR_B1)s#rJ;YX}|5!htPbgC&h9V_V9_IBg;`u&tvU?SGvzZ zavAH z+^u}%UhM0?TeGbBId$6VbP`l^UrqM!{KhG0ZLOypazf5=eUFRKV>Xh`)oGU-Oj14T zu%pSI$(IIjQ+x=5HMg--t??WOQublC{2`_kuVBHOZzZz-T zzixCW*|El$GWF;hQ#^5J%7@mWs~gVGf;UZWl{6h6%t#9!TE+$1%Z&ItqyVi0z847b zehyGsl5|+0?WBZgcezS5tgiMuMG>9f>UsISC7vqTvRM>vxNy?F`dg$>PxzGH>f~Kl z?j!G6dOt=KJXG9A92YZ|(;3KRaLCA2Sa&~<%(lF}WVZzINfJ@mkVyeU)$}aq?LMq5@3>aBs_wQvc?I>xr!&CpdN(f*l;j z+C4wm5or1SVafk+*V#4=!$u zIz5s}PLD>3xRK$a4Oy68O7@i|=b zVO+8G9X&bKHa)37- z$x(am3k8E~BK-{;G7#k%vg%^SbSUIE6$S@IPXp(`g}TxQ+Ix!Ik9URxl5EGFrB6}H zxC38kUJPSA%*y5X5Yk2-XB=hM%&2Cd#Yn9Uc^m8t$ z-AvWZy;jcSX=R^;_$#-1ekSLUdvJoJ+x2S5E(u-Tw}dDq1fCqKJff{4LY}%lkZ>~g(YF_=V1?_ zTQYs8i-^7|^d`00sXw_7P*Qte3##ODyIHmVvR4R9(KzT1W}(>Eu>bx1_kNp#pKkhQ zaiEBN_4bGzdkXgI&9kt~nv7dXT4Q*D_N@>&x*S579hVik?@k_K$r9IGY@D-s**+gX zdvVIXcLK<0VE3q6cu^Q8@;-}0PrTX{z`BT*g#$*Mq=&{kXY*D4ZM=+0R$2{wt$*jf9ULtuCIN)3+wOiE0*)V1*2k6W!)E8HJ5|7pMM(Sc*RqgveX#&)4!_ zriXa%g4-SQ_5Kd~e`q=^nWkn_zZ}lxQBdPVk1X#j_W9NOXyV`Y>^P)aWbQ%x6~S-^ zJ`op@Gzb-t`2{YWs`4`H&Vlit9IgLgvWiXzwb-5WuB{%C4;4UM@}MF!mNo_`Rsh-tOd~{A*PjGGdST1VBE0Nfu3J|K)2Mz zOdm}PyG@l{j{r$+zkuIbkt~x6D^;$eI(-nZ8trgl=}wozIU6oLnKImmFY{xZWd_#K z!ZbjC!83%V*9G0cGO^EDVmo5T=Aso7O^DZko#YaW`_P^(uRI7VZn2a6lVfLi2FK2T z?-yRvZda@Yt(aHFfNTtO?zj1bwNc8&;H`4`^|n_s!ve4&#FUZ|%mL3F=@F{rzMS#( zN_TRZO1Wep3-*DhEz@&ohcGXRxa>y;7;Q2Wq)gnv7E56V*2r#*2xk$Y`+527h`I24 zuf7$U%6IJZE}PCfo=f`V)`vZe+%(cj|8$vIlV0M{B=cMLr4OTNPW`7W9^W8oxI;Z3 z#xuAK!nw~v`+~j4qHsz5k*GJ#$q$U+(iA^9YJ8>d`?!YZPxZ}dRWJaQG^TnKj}C7i z@*{o+=qdhO4jdb=a;9sKXK(o$BmgXV5TOjfR?2&Oi>O{47z3?1>IH?;SW`n8ck(lY zJyxZU6?$D~FhqT0W$om~oy%ak!MjU^3CjvaUvOu~mVGy_%K>84xo*VGzL6dI;d2&- zYm_~XCzIe*EIMO=Md?L88`MnF$7WaiO@&&ip2U1mE;4QdnHTq22qcPLUBD2(9iwCL zi|ns*+&3FI$Dg05`~4lxMzlZmpDn_2@Iy=RTs5=P=FNw&W=eW#Kj4CvQN)JoJ`Q_~ zdx<~n>j(MUhn)-|H3T`hu6>Vz2ZEKjpVI>CS{L^@X2V~Lr$H}E`}7lSVb$}E$qG@a z{)J86{O}o(%B(TrwyaxK>bM{J*gm}er@Phtu}??b^KZR6^aR*Js#^{0DGtfTFcD1# zXbcpu_17>FNNv#O{~A>Yx3c8?eEd?=y=q{ma=#0=a|zWtj+Ud2ecwOVC?r|9?|5kL z?~|l$vcY+(v1;SEzpZ@P_S<4~P9DDn8aVz6?GJ_PZ!n3u-t}JX?2kZ;>}K&{xef(h z!;63CU#2}R?qZ$&dhByO8b8sXOKQq>rZ_Uw&b0MzUkz&7l;_3Mad}Oy{A|MXqJ}Y> z+it_Vs#6XVhNvX4S2d*2^Y{LfTkeW;#C)ddIuSos3&zcsEL_qs@ z^L#s6i}~t=&XpVXW!M&4N%s}AQ6s-!s1m3wMfK3i>BPz!8e;s4d|z|Dx`;Xb03gZs z#PehSOz&slq24G}`^RC(83&UjMmxiMHL8#(!?(ES&<-lhz0OO(nt$Ib`WmrukM|WH zSK@DAHd@XhGSJkxI;)UBS|DstSf(W4zKDp#Gx^<%a~~mpsWZ40Nfxj~c!&z{$J!e5 zeNBjrw@wyo>UTJq=Jsay=aY}4{2oS6U|)|P|0oNCdFR=aY?^2K$pY|Zi+8;)GSh(R zk^ef&zj&bX+v?K>bGx9%8ApcyIG=lYMAIRnYpH#1UlD&pSIdN%!oag5|I~?;dp3rX zq!->`76A{TM_T$@G_La$5`GyR!sq%w;4pXz% zbtY5%Vb`*qNCP+2JK8UQfK4LtTRzB{N;MO1>)?;;as*o+%mfZ5y`A|}hASziuG1xj z`gy$0z}qP$}J00gl12 zeW8NS?=qMZ={o$xeWGWbD2D;+$6NZk=4debaxJqE6QxpF#AfU>vkh?JW=6oggru9H z2DUQwr(1riy!mWtnj>IEe~2ISq_s0Fotwsu(VuE&bK6fzsg*BmXYg7Zw+rNgp0cMY zuN-X*nZ0FKKVu%i@j2s5JvZQ-;F%z&n8?lWPXCJ~7NhI@py>TF?l&A}KA{7R5kE@w z>&b{;M3Pc^DtQn+_CT=-kADP|MQTRo5l;EZ7l`)0ucc^qzBU|EIAqc#!ELw!6cU|5 zKyv6iOn}EFNFSE~KIK{r6Qu3#mC^C~=bi~K5PQEQ6WtU%#BF&1 zQ>xbqnmM#Z-AoQmJ4Ec67h=nOvVz?TZnI7HV1EX-DY=nR^oFdx)Lj1uvlrpu8{o(HeSPPEm5AYyb z<2!CL3BQ$SG#$syQpDSJ&xyj3G3{gb=l=EWKsQkQO;FUQhyPXoMGiHBS1RnG2Js4f zMys`Zim`=zAgiOPgbks)`fB+B4pSTUfDA#=pCCuyy8HP3!4ssL-Xo)5VGA|Y^drTq zPf&y{_6OOol^tQVgh7AD51dRNrNrVBh zQv`3PdANUR8vX>K1zlS>{PM_nKj41rwJ?-k;*A7eiihCPlF+vnwx8y)ws0COEDze6 z(k8@iGk7P?0urLr08rhP#h&Q?F=M54lUT(~dD6C{Q;tFVxH+)6Ic+psg-}h03oFNN zILuV_($CC}|K}{)C6w>&QudS4zoa>Pt4m`ST$IGa)>M;dG#4lBidH{EyEDunzU#~&Nhlw`q9g=S%VhhTYp($& zv~d$(fnbAa2T3~Mkhqu(TzTC&Upy5vF z$a`+V%EtR`I@5r}p!EBs5^uLqZUK2jAM{b4QQ-UWn7U+d%x3K0o4Txs@swN+CcXvI9z4^s$5w8r6iwFd} zDNvYqjV=hOm}BrrfdD=oR7?1{K?B8YP$h(Ppfp=WgQ}+xr z8{r|b$8Pi6_E{rU_wQK_-)S2F(o`IqF~}8SxF*p}e>~W$zQVMTK?8K2F9C^c6hgV5 ztA`iEy`(=>Ez_ZHg8GfV5yD(hGdfBo`=maHg9`>m=MW}X$m1t;ph#~>i*z1-b@GZ^ zIl^NUIho^BCn5czNHo9(ywkE2IKg#t%3P7ThlLMP`I*5doI-8xp= z_1{tEb$1v_gVgc}pp3lAlh4hIv|ZU=8O=3Y8l76W1r(@`SSXi9|O zA``^f={&YS@e4*Ck`;Kv-&v*5nA$40?XkQo9yWJPzt+PXAcO_8t%$%)bF?vr=zI_F zKNiN|6m|rGFH2X9s!=2~rBlKdcwF0J^c&@oKipohW5vFlXVQ`B-%K3@a(g=~hwd}% z1DQ;{2N#~yeAxg!{QBz#x8Veq92ET03@I6T10x?D2WEt4`=UJ>M;%UW?3+IUc-8w* zScod>uKR77@ZO#JO6mCOW$Cvl9=wq=d;{CreQ&?=fJ7OueDX(V4v?l!bs#R9x37mu zdbGyW6YTUnvgnJQ&DR?@rfi=h*L8G?I6ybLKm^}Nq7 zR*VTU%&%MniFQzlqS&`D!)V1f>_n&Nd!>Mgl=HnQkAI9!>0wW}ORdTKoNp(2!HX zFZIz2{io={N$C5n&yRJv-6-IRdKlKW3>|=<@(KCFLM`1fpED$b8^J|-x$e^1!3~d< zY8j^N5MJBzZ2w#wGP+S$(1kKwDUdCkgVUMpC3m!C?Oz6{c3FtD%!nxJZLlZc-{#9A zk-fQxwJQDgAY=(p5H(}y4 zea03go7vpKB-yw0%A!J4x1wi%&(K`O$lMT8KF`zRDLcdzFMX*OGXkOZn-^q{=QM%V zVDgY$d$a>7kIHZdi7#{L?0KLUDVI|gf|3|YfAK+%^=bu8^Wwb ziDV^HWaHbi{4!patiol~;><4npU zr>Uur7mUiGQf|;4s+!2XGASSjtW=mdIU09O0EY|(c-cP!k8LvjP58{)JY+O983ML)WSnV$d`b?WP&O#wS;NqIaU z%z}m6_=O=Hz;Uz4(RRl4T_uD8%h96UkT&^dWmr7@AyFRi!sdMmf=W^U01Ouq>yAUi zp`5NGUQ-_|wWD5W06(rM^<#TM=IACsu1$hj`huBGULX)YlzMv83?5RPe3PREx4vLl zZ3flPyb9Enxv1{Y*5QvASY?mgY&dsYSeHs}{kU7zT^aM-<)ggOLu85LiJsW<4Q;Vl zjLcG=;39g`fm#R|b-IVEqg5{RA+_5VF1b7%VgrB^QvFXy^4~KFb4?kpwwmFdjD%r0 z{0{bRC&nVJW2%mZuF#({PUjn}(XxGTZ}rDXIwV=6vDLnVW48N9x7ysCDw*3pcUXWZ z_7MuB=H}b24W`fcH)_YrPce&!I~<@Y0rVcQbY{={UVyF^F+p-DBf$+~yq^w=IcH0+FtF5`#%A0fMj%G6Z7-5^Bm{y9fR3N=ar?y++CKEBX&Xmh*dlBZ7noOQco%DW z2#C9Z;pH(&xvxV7OlmYVyhZQf-?xo>!g|Ar#b|bn z5W|Wy*=S}Xoljw8SSAWfD1eg<%_zUlwUfnAewBbq%75^P7(ev#Lwd= z)@jdAceM^^g|u!N2r~P;q^Ebp&OIt}ud2LE?9f>3>qH`nQBTi1DjAbU4GGpBocBS2 zqTKSv2Cf-khe0*Ft69Mbc*Y#{hznd1IsGeDhXbLyhw=)1Ov0SLr+u#MDRWi#HB=s- zneI?dTvL{RPhZF)?2%E=eqe&v_uuAC@szFD;&rSQ>iU962cdVIBfvvF=a?>YY)Ttf^$`K%&?b$M!U!+~KAUS@1Lpw)D~ z|FSxVkCB3-ru(?NeERC(=Z%NmZt8kN17=ow@P%}~1O_BRNXlhoq*T)|zLDKoeV^G} zDTw>L+M;K!8ImqE6gVU0mLzJ$@b}b^Q*XVKv|1>t$s>O+R{)2i-VYck$9v25CbD{e z1TASYTSg_nFOhgKmfuoFR%H5_JO&m~kIeT|XxZfjXF7id)G>MoM>caF48VTF59|n` zMfJ~cO5>bZx7!KN;v~R104t3@_SL-!dv@Z6UKh8b#&qLOI|)bfx?d3c@Te8Y01Ee@ zxg5{wi6gz4pRjEZAK#O5b{tUKch8)f%_!KRRK+KL5v*1pRg?op%k;k6h2$es?dbv) z?qDvQ0#Vkt-gEjb@5t}sll0JYzbAz+^^CFkZS9ZA7N?I#?{DG=`%vBU@Ypd% z7E?@;skozhymI`NQSCL&>(u2*;mkhVNpxHFpf?)1|r()-~x-U@AT4>kebxk8enfEbyYKor+&_ky%bya14G5$s$4LcVb->{O7>jj%w zctlJFzI+L+J%Q5%x4go%5Zv{;^sDi0N3A(t#8kNE3dIuBW`3#W{fOT_rZQO}d=RvR z#Q1AYA1|mYq;hamZUwSg6!72SudUvF0)LY);$ZfKu$}$BEAd=t`j_gXWYEl3oPV%=U-ZsouqZzMT4IQ2`)xi7cYMCVGI+5lCM@ynAMk|`?@Oj# zc4W{VXXYZT7!*04gBZYj#%HC;9_fzwPVT#k1>#N~Z85)!cLEJbv2PzofIuYHTfmAN zH2-qaS=$qi5w>td?BplhZ9cEE_tIUj^Y1k5y9C@1&0Bu;E5%oiqdvYcJni=@Bagp? zPJ1m;yh8~`Gf`;Xvy+|LwfwnRLHdzfVAIdb&D!*p`xgn3(WI0gSMg8N>8YQ9jLsQx zzzQTQpI?V<$7e)wcX`KON?B|IP9wJ=u(kF|vL2Foww>moqmEMvj3fDYBkP+A-uYD? zzzzA#AB`yQFO76)ggYL&01Q7HhU)hdNjw~i&oD0^api7Oh13C^eWEgs3hcuDd5^iS zTX?vw@kh<~J%&4dz~aQc`axVLGzY_coisE3Fqgw929XuO>jk^NcBd<@+;q1bavwxK z@K$=#GQ|r~m3($}Sd5FMeKPfLFM5g!r&KU~Mv11@SdNd_{&fB{^K_Cezk4-%d&k=; zdQ;ZE;0Iv&Uh*r@2CDmA=K0SZb##!i4Ti=D_^SpJ+0i06Zd!!l_;j`ETsGgD;T=Tv zB9f0x1iX$AG~X*k$W?80*B6*i_0XUwEQi89JIzSQ$LMpo!zj4jQv{w!H?@N>b9 z60u=M^5ihC`wsgrTB4dmS8)nF#NXM+%}p(R?n-vQJ|F>L6N#?0m;fgp8X5L5#X0 zn3vSFX`9zXd<2`zswDmB2)^VEw{V9W;Og(a3sQbV$jMX)lDm)7Zcr!e-}BpesN8|% z5mk4i_{D-cO)Om2ZzZ`M^}R025#=tI4^kpOO-J`2A&5X$RnTb#i+Z)ZKi7WAp0Fmq z)tU2yJffZ6*@L+DuhME!=%Z)6k2`mb@-L+j!A|a*4wmdtgnUBbJ*dE9ZY2Cf^)GQd>56z zsq8B$o4sf}n&jSeEEugWSR@be(ZN<=-AzuJVV{i1W^_mS>_gPG*pqTE8%eq5?$~j} zb`*sR1$a*ut5MHB4w&%Gpmp=f%|2zh)9S(FJe%Z z*UOuBI{BXJp##olxE7FPB_~T8ENQ1wu|wx!XiBh3kFJ%HSu|B$FB|Glgtt&i%sjZI z!KN2jX*)m2(T8WQqk;9JOfUBWJ3dDu?Cum42sa`wMY;eh$<5hkGM5DlkK?-JJ~jzi zO@8;Y0VCbljQarr5mW;4!u~;F&ad(P)sDjv$+m{NBbRS0uPvQ&6f#F+Xbp?WZm$&$nMYV`Ku(KcLQp}#5 zbRGQ8%hI|MN%ghR@S);jp%Om#7Fg`G${l?3$F^@zU+(S%iJv1w(+|goH{E??TAAnhH%^*4=nsKXJ`U zJn?xGMs_J%{a)dfzsVIQbzRAn!i_VM=+uuaK#-;M)q>mADU+*k&Wj*Kr@oM+Mi7Hw zaX&kVRhu8_myNXIpKz4PfjxLo zlOcYO&$uXr9|9bVQahUbl*H(?)GH zM;HJ4J6~Z%VLk?8Vtbi{KS+nP?PCU`9M zP8E@Z;vTQj{t@@8opcnhL+bzG4u@?_mY4NBW+`Qj*|I-}Jg!iIHSylc1?vnBh<+st zU?6$=`Pke!O+w{&%$Wi*`Tg1DRWF;;9)98jG6f}PfL%%fH7X~>Gn!ZoAGxp=cNOm3 zijeflQ+L>@@8^y_o>jWNU?#6o$IS1voJ8{3w9dtIzWG0r%-kU<_a>-TAj}qbzq6&nJ(<^}l)dD&pe8^}y^r133H~Uhd2<3LfPuPl>s}`qW1@zCw`k~B zu9P?QYNM`4Vr@-_$O1EwTH8PuvE%1%kM(h1BITN{CZc$xKWXmZGenUS5r}7Znw))m zB0zVAMbreIdt8q0^LC)3@KW=QFZFOF<=NQK`GWvaPTOBSxTNsLx!eKRDeB6#K;WAmuyvwnA z@{l%pUWJb%-4lO|;@CJluSaX0WBSd-`g2+teTY=&Sw-SkO<{afc=yTge3N9Y|fuShxm zbBO!XytC+XWnhkn2auk0uldyO_rNE2m9Hlot{0<@=L<|O5f)u;e-g4ITtc^=lK_H> zG=Lq=8kF(E*c|J!-xbfVfac|8AnI;)FXzCL>&so2$T-28<5$Dacy2dj!qVf+zNcZ1 zZ>A>Fa1@4bo_fYjDI5VXe8X7`02dRNt#HmM`B0=c+m+j~fL)GvhRTe)r^2{4>5 zUtTMLigKbx4hciHmDpR;(4(-MTqucf>$ETa-TLb*Ak2eI+ zlpqV%G)0|6;^q<5tVc=X=+KfffC zO?Gp4d&k}FF7AC!xlATgzrSz)K2J!B^C_1d!?ZQp64lkB5y>lc{Do9dk?W@HJwhW4 za;{uf;qQb!6(x_st3W4G5FLhm-V-U8vbkV6&iDbA<`ncPmy`j78DJ|D>JT9M^m!3= zD*|&C@H`&X;6*MGWsB_bjrx3AO&29{08EyHX$JRt8i$t|bF~fSybjepY*y0g`z2WmB zf=U-9ZQ7U)Xx$)n#u2P(7FlwqK`4Rv$pK2v?o~L=5e9dobBQe<+QU8QfV0N4g|M~h zHlz2_mlGqIbhART$!s9Tn#`IQFyak4MJs|@G|>}^qDrbnZ4P}>kq=heH61}<23$Wp z6rHJ4ZlIMzne=CBN+Ht5Z6q_EvXg4~nzTI|NThs4%3*gV%5CJ-D0TO-MVV%&R0q|e z)}qdN!wNV?wVXE)wb(OYXH*J}vYCyV0+iSW2U0*0@B_jEz{}=htpX+s;OKW?A48O} z1k*rKRufRnnj}^Ov1T-43dPi_Oq)0l-G0(QsOoHxhZw6#X7HRM>(um!;dILC62Xy| zR@+(?)@yh0EZAJ}1Vcr21Otm!qZ59PuG~GY-L1!rPtKal(@`Q2nKr&sk$)5s_j+@U zVw^|0J)TI=Evv=O#?@7|s)p}Pp%T!bI|gUE41HwEg=-$cM1&u+&2Fgt7Fks-Kp5&V zR7wHmG3F{7FqZ3V)bRytqjGtpQp+h6B`lVv}K3JxCO}nyd%dU22+e z<>}y!;dq{E7aKeuBmPOW7{&Yd6p0uZIlDLQ>IA$_OAe!YX>gB|T?*X+P95AAbde+b zqG0h0hnOHjt2L9-TP2G9biRVaRwM2orX}%_JOQoMMj0>x(iXK*jgK+~m4MQWV%E3? zijp82CA!okT0IABFA^AOb(LBp7xycJbdjoi;+aHR^wvy5NR!Ql16}5&E`D+>oy6YF zfG-h6i79U9TLD&*Kph!!rqkp?o7yg#N3(#$auy1^EB<91$)hy$qD_rEP~#fO?P*PF zFkJ@%Uf_5U9cv?eY9`#FJ)$PshcLX6ZVY&}9(hv$`&FU{ZgJE2hW3#-7oXDj0;&BX zji1FjC8ameLfwyTP*}q{MiX{XN?*>Ni=rf~^_WrwH5%})F*as2ChFl}CSicSLC>SJ zlQIEby6rP&bsmj6R0uayAWE>~837Qx5{p2Kt0}T76U6tzNCVJU@Lh~81X&&Jt1;*n zqpIZEFb7$v5*4B~w8+&p;CNwAqhxC&O;myjQDCi7#v;lG_&qu~uAl0E%jBqF}EC&GsaKUc4S3w0G5K>I9qqjNTLXNv&Y6+|~d%hUe|tnoDHT z{=6XtY%7-Y=^J`UN~t-$3nxQIyowb}ad2>%Ymy<`AWlmu(`-g~vF#~DsSJjq+zreT zWEi!htgF*murn(hEgvy@yre@1H%PxjCf0&qV!khv%2z~xFsS77NdAFHQ(dcwW|r3{ zU80Eo(Rc*bHP;Aeqf{dwDFh@mRyPf z6pa6Onjjm~H#AO18u3hl_yPj@Z^~eQS2{1i1s>+xJ2@uMRMnomQpiLkj?%&Qd&rAuE>1-$;Wr>p~wRyEYc zDl?AH0UYK7LHkNZM6x))jb zu*wz77|bzrc>{*L9R|m$xvk2^-Fl0~Q|Wl6riXW>)J`s4HTpA6d#SBR7W@P5FwhhL@ti*IY)i5{SS~ zU4p_0hNnQ%l3BmPQZ6!ZawF*CKtzFa~AZORycnx z4jLxLRaIt1w{5+SBn_s$irSf#(}B=&N*|W@xlvc+6VjE3s+8&ROcw zq`^heZIiOBp$s`2jmn!jBG_vfc6Qq}s+1>ZTy2qFgD$v_#F1mX!LP-9 zl*ekV2!4yl6~#N(^Q+!4_lUYi+GsXt)~LMwK>~6eKv)Y4AU_ z0LqtEqyiLRBiI82`Vh+d6#5bf5QNoTrpzX{uhnICL!md@A~A*!)QWMV+Hb3C@zZjM zirc`GWw6#m(Wq7Ixa|yFaJG>(Bz*C1-nd$rikEAviNZ zp9r;N$!|r6LLO?ZP|D|R%9y}YNbF_ZN_CLWo2)UMR9PT#N!gRChbrwn8cT9r>?@1M z;6~a=2jF{Kyc(1!Pp4x{bh8v_ec?5FwUv*s=}geu6#QxMt-76g5_T4V9F3tl-V<%n zQpH$9$wI5O0Ixh|%m^tyl&My+qZIM!^+H0StcSC}{n4lBnnsH(TCNkEpa+_ikf%#Dic@;Q zr(`>!NDK_>?WCjto;>A?BZUfTkDM!L0ndgM1!^2uLsw}lXjJP-?95U{mSb!MOcn%M zt_+#0izg6p=4+lJn2eJlvz7o1G6)Cg| zpyH8P(3#_n0E^~&$yUuris7cm>hW4?fiUKHf=LzYQHMzOI~Q&@!mUKrY3-!dQa1$= ziK*SroX7>+Dn-s8=ge-D^s-i~&c^}}iHnxe{laIxm_zrLkajc!6T6UCFiEpRM3a1n zDq2+veSp~Iu-8y*IHZW$!-OHWxne+>W+Uj;QzZiLFias4J_4^SuFV~0c*xt zs0Es=CBRXNbjE=fRR9h7ZeIx)7F401bc5K+FNoN+t#GF zfIdMV=S=6C*kG zrpX?#XX?sMI?FUGFlOM?RfBV`60fEKm&R#P5|7l2dCVm6j*22#;eA=F!a^|Obd?I) zT0(Iy5VggdR#V8=P9qVb1OA*^*-m9DEuoCg*?K^o^^%9`5bUTBX%gJ7C#p6dRUcTLRaw4uJn1w>6~W%ONwk zJTS&iyDI2Dat7%zTh0VJL4b*c%%ahvtp<{q8FH#R6w-|7+*PBSwXU>=Q^qJ+&rO|@ zS9zmxv~ZwQoR?zxVm48+X4@4D-ynV-&`s@z5KBp78629ysFK5E16`rOj#e#rsyWQW z5P&linCSIx13H7;b&EOV()lG-(Ik{yCR4%(Y_d$!Y}V`TYL03&HN+lSL}k*aU~B0A zN;s2v?If8cMJjPcWI%#BF(J-qW#IqDfu}&P%qE#UO;;2jDfg z(~7yPqG+yjY_ZVPyG;RCvYqhx91wM-%HPm|sXhgA-E_JWW8ADaqjsp!(883#OH>Er zg2n4XXA}eKO$yw%4J7(?8oS1~7F$V|az5mxvsW}z(P zv}n0E`2vxu6~JR=eXi;b`GbG7fx0hR zahWnkV=SlvoFCo744caW8~}eJlV$>8)YSB#moX(YRc9T1vjNPQ*5X1Lq!mR?hcD>b zV6A2;SGl6M(q@diiMBebtEq@1?L_IfiiQ^A1;K6WRBY{b%%-WkfsGtW#3N!|l`PPn z0EV#bzDlF2(^))P6~HurSLSP2-R?So@9>Ol&XBgvFabtInJLuZDvYI}3RdGK3{@A} z6<@boLeU7bEbUe{7~Y`AkHS(l5^gEoG!3>gQ0VF+g&?tS>Zm^3D43GIatKsq=~NpY zu*L=lQ)e%Q(l%SE1R_HaJ|%%8S}|Fvjva$7#BBuhXPm}@ssf4A+iY}P;kZ`iE}M~p zH$|di9Le`gNX5Ht`G7wVz$|N(H@A^azyzDB2nN2YV6_V+N>52f2k1UpnIs+6XP~5Y zEmuuz*4n+I#txDrY_OWM;jCTYoE?K?bW)uVoMEXIi`$6-&5@=`6B@r;UP@QfKoMl1 zQpg>Hsvh6B+d?R>XFW_VQg9RTGg?{(r{7ff>9Z;yNJ8T9MOBS9xcGpHMea=9az#T< ze_5kV*XtciBpOn+W93|{;kNsn1vvQ`A0T6GRnDLB7E%pNIu~3Lqt~=^hz^`Q*rGwY z%EBXZVhAl{X;aah*PhXu;@&KB7tsh6)9AxyiHnO)D;ngZ45>q80bs92$|4s*D@sQd z2NbX&TL#{69r$e`io~Fr0#>!dVN>jgkmAgm%2S{V#52ic zPT@4FY8@NXLgzw|2L`XjBjp;v5HL!4fOHy*Rp15S0E9~z>c*jWXSujB5f3|*r1>@% zbhwHUcS4Qf!>Fmi5TVp2a&*;FCAS4-gDHnhr<6)qJf4UY0_2TI0~AYT&Vp_b*e4yE zR-eE)O43d@427nH2g_C|#(>H#nNdNfDr-K`R7w(Cn~M?my_~FPG~txgc6$Jlqft8I z)^yaUz&C+dxl#_rY}zKY0u>5|H2!?b3D>MuF;^pAYdi$nJnYv24tW#>Ja;GM4i_mt z*-nEvGXl&jM=PK-Q6_L3iD`X{vN=;_JMJ5OLbxr+tIq5eeR(ixp@0^%inK*r(Q8wb zhel^Ho*tSk5z4!za3ZG0&dt0X8dYc0)h#GpdKp>)?pBO2cyp$7*qD#T(6Fk>0imd6 zRyYETqU{IDxvyhuTY^y!OXWgka;|#BQ&0J=E@d@V2f!>&rS#LOf+x=gbHt10=?nmj z^`gZLEEJw^vuH3=X81htFkDGvF)szuKEMNnapS3?KVb3{Dh!{aTuX@>Ta^g%qoj79>lbIyx3|3hJ<82(#}m=}EeG=fY~>Jxt)zx% zBzhbqZP^Cf@IX1RZbMjdqbVjIvbK4JTT^hGtyLo|Fu>ZZ9;48983V?45_k_%OtPV8 zJ|CqhyPvP;ihNDwY<0XQutMr=8ZiCE!ZmL@`caCw9t@n-_+~Pc5whNz(@dPcWDtld04fZH3#x{fQ@dQ?d`@7b z3s@nn$Jz;j@X87h358Mw|D}Lw+gt+{!VTO5MYG`tjxf(#Y8A$3$eFzX3PZm{{be`< zRrSV0W?ddLZslBEUu#+HFwBy|-!^5;`G`jy;p-w&lR;a=!M3w()2^iH3?IWdi_XSc zO(`NQmLx`VqpUJWqLsL!1;ETIS_=}uSC0EE2BA{%NeP9RZZ-w2KU`=z;tkwW%|*0q zDB?pSW7$*4XX_!0Mx_jd!P=QJ&b?{ABB%|O#!H>tfTohG2LfOpiXsk!rlp#lsoTfId{ ztpXTKbaxQT$J3^gzeW=BTBDzw#$ZaI+YWFQtdQ`bm(8wCh9LCHOx;~@T7#NEAjAhE z4Rg(iQF0sKMqs9_sKsbhpU?W8$fEgjjRe(cgu_JpNw%`KX|Yw!VNJ-8JK6D0k+$xKk3dP{m0Z*8%*8OIs zt;1=9s6YS^*~u^Dm(mbuMe}(R4j-d?4p7x3l(I*JnpoLFwYK?3Z)?NYm#%?NGI92B1apKEI>t* z;X>O}48wkgN|;GDTWbP_TtEmL($p+2t2tN+2jlKgC*~pzsfkt_$T;}~v^MM=r~9ih9Qu^cPWOw0U0FF+xQf%6LDF`r?gFz|m)7OFlqg|;I1VbDQNQ#1y%j9wizRY00j}}CV=E|-r zqI_G`r6}ZL?G9>$m=IUAlldqH%AE~OE#**o0+bTC*GvrtN@2rTsSV_trKVb9>M@s> z3I}6+2+3QDDWHPYM8FJ-(P3%{W!06JS}8v5Ox57=Bn!qmvKF|0^vZZLuDDVqMFPkn zn;UgIW=Qhri1mU8xGUXeiB08`rjCc`cP+a6DvT+bZq})0o37OhZOFY2e5&|=D^U+B z#+1r;TDpSW2{1%^KEyf_{&vdf5di<1_q95)7FSM5u~4Yp4z@%rxGu7+WYq?AN-LL= z0)SAABKtWYj zNfr?Sy3Uaho0*6kg9|FjjL#k=sH25=n2a6+oJ42wY3q71B_%XEPagUnXevQngRcQ% zT?$84wXiDT^JbutDoubrh<-+85`dB@`E-25$Jd#1By5e@vnbhsfYgo-REG8`n40A0 zIyDaEHfs@Sp}OB+Ewk0U%7hF7(7nQuD9I*S>%5rNhFFwR1&`Hb&a|M4z-@pUqRr}S z+T-q0!LL*WkzlZT1E|9!J?OOcVc@9CLKV(siCdU)_~{Jh8`-#7tMdvuIOh#VLt8KE zHB^?<{U?sKrl`&CXL=z$;Bh0F3F57THB#7vz;TeYAg777&9 zDGA{2=|};j&`qza7|Ui-REh_7s!H+$`5wr5cF|KdTDg+MWM~Drmm-p>m96L+aH6G1 zE?sZhGqs?)6i}gX6U_wy^sXu+T#vvEqn|HS^oT)L%Ep|5oJL(1#kj840-v^$VwLqY z__ITiYAu6?MwRHT0C5D^1Rg`#EV}#}P1yzvJaG9f z18$E{BC3V=7t&Xek{#%uH!`w<26LIh=H8AXhO00QZY4<2 z+)97i-*6fo)+TWBd@Y5&q^tOgtjF6Z)N=|4TjE56ONT5z+W*MHDWv>nt)UUlCEJ+G zLtZQfOk|5Iriu#*eVwla6EtvJ%7Q3yAtWk{l}^a*_Qchlki(Dw3JBn1Y;IKBO)g+` z*qyYWkC_D#Vwpr#(za(2(Zsz1P6V7-p$HFFSGDs+ZBSp1hIPma72_zmlwgcZZl~b& zW%99Vk~OJCtHl*6muwux8#8*=6?f7;Uk*emW-$>^1&U=QeB)@a4ef8MgzKSHlVd$a zbCpwuYFf%=MgHHV)vGle!Y7PUpqr2ex<mnOlZzl5rGT9$P9BD;7wdEFKoZ&SVRS z&Ll)|hB`HeDi4A{_Bvb?k*`1P@F~u8CpZMx^LPoex?A1x+2#_&CX<&;k10f})1n1+`AAbf~Gx!mm}7-O-8`SA`Ne{9uO7)5%V(>55ugzzlA;>X@-t zQA(W4W~}Hurcl6Gai!uNe-43c!Qxd#Lio;X;GrvZCVez$ZCUjRrKsbrKDCk+T|%tf zs8e=LQ)MVt198R`^#p?Xg1wdqnUsJe7s@)T7g8oU5)OF$=;&zMxEf$g`J%n(DON)o z!2Z-}hfpcR>}`|19m)q|;WlmeMd?-%Z0L~=u0ZFu#&v;$Bk#Ax6LFKvCX{qluRmLj zC0p5)FXZ=-q4ffwHL78ss6<1K+g`{a_J}E~QD2-@3t%*B0WPx|06(s;E%Don^dd10 zQZ#3FDkw)Wq|}CNtO}zm#!4%uvAgURg}SSdY+mp@;3+sXVUq|X6kme@mYltlDm|&<=uSfFrDk!kcI+6)(0_sBwwwz#?X3)A+sKsh1u;`LGVCClm zu5j9yOi1yT--}FCqY@0-T_u~vjJl`aZ8ibAy+K!F_GAf445S()^?@OElzDAH5UuJm zAfL0etZHvX7nhVB8f zj8>*d)m<-H0*NXYZkFgg0OvSBUrBk&AIjC+ywq}nNJA{uRRv$3ZnGu}&?4;^NNLe% zAgih?wiX#W;e&O-W(>9gd8uuh`E(o_gs&ActjmoB;S#cDLxZCnE*mD4ymhn7+jcZO zrkG#D`AVRQ5e=bqQPc-A23@tDj~RjiTPIK~)LN9o7PdJuDTGuZgTaso!+F`8Eb0>q z8nZw!bDj`qHfUI?sT8tJkH4l7XwY#4^EnTaLt&r``XXMmnn~z@m&9zjk|@DV3yXDA z%!H&3U)JfH9gM3MY#k$G6#*fN$5dsUeBM&eIqI=wM{5ATS-lv^xdL1^ZZ;#2P%l<( zR+o)0D-}#iDTXOUO)v#QZn9<_#dus%HQR$esaWdxk$87YK(xXDo;9Rb)H2{GPXTdQ zk~)TxDVh}Y3Qmw{jl&U-S}MVi0|`r)!IAO-MLgE(MEocRN3|Vvnq*a>in$)bP)Rc6 zNH#3I9iacpvAK{&5Nd{Qzd+QCu%zQmXt<8ZSvr_D6wMBe#wqw?gsKdq z0pL_O(_#%g8U-4&KYH-(8PrWytObx(tg^9sA<)dX(9D*w3qTd9fDaM%#td6!V&-Zt z#sTZ0Xbv>Gjp|x`Go;}}RGQH0SP26C4!mno%CGbY7K2@1@-l!J;>}Gi6wycX)lw~z z#F!|0n(Dq#(&EO;WbzMoN?f1jyR{7eyhN zB$urNa-5X%S9*8d6!{9aiRF3;a5DJ4qWekWk#;4AqX9Wgo)|elIbsSohDSyBA9IBl zalWBVjBz&wc?7Ha%jk$$ z;wE3DELBqlh?*uqYb>Ut(ljJWANo>yNA^oRYC>WX_r@2A3;|w+{71gm8Hsi8TGsu= zt%HkN$@M3N2E=M@-r)LsA=70;@|ukrZNo1h1WT#gOWg@295Gqi@UN~S)YdaaEZ^>0 zUo&+y&|KZeA>SF1`pGBrkyKI)s*2<`+kK6sW~XZ@GNw=V8P-y4|95joOZ6{uWM1sU zF30J1Z@Or%#<~uaOB}{(%~Z>Jv*(m%nmX=Xra?%fg>tYM7R)$JFQGeE~y}XrHl=x!b*%VW#0ZFju#1SVe zPj*(DanO?56h+X{I(>;S3y>OkUi8Qzd!BpOL!Z1DT9FaugA5#hpri~(0%Q-M9?2W~ z3p`U7l8AJMca+NIf+SQ%Hzvv@^d}7m610K`v}tp6w{*GCcf5x#!;c_GK5g%YD+sT` zlVJnL_OFI)w+w|2N?AUNyc}Zm_pD|4WVzY9?(U}P?@WGDD|w~8ujQSl1r9B`ZZj}Z zmk#+(DQb^~#A0Ph|Ku}?S!Q;?!Bd)y($@Kj+ier2%eczgP9*4wqKeQEz4D!ms7@k^ zL*5}0TE%!B?{{oKIuzNag+f7JE|d|}BJyMu3XDIf@H1jPJwcQe5jp!5QBI-(B%Q$am5XfmdZ*r1PaIOo zDJpn48M98Mct}2#eH}F{pYk>ln1qth7}dq&gmxwiX;O1yXs4T}>*tBo3grab7eMmr z@h})pHK9~16@x$h*hW5gty~uqss3@uN?N^!ow}x{Tm(fc6cZe+nf%jxC-Ma`)C8v) zx}Yg57;=?KNgv{ZAe20DeB?WN&{k{q z@Ooz)cNS{;SrqI3ESkQ+_HTsJ^lxk(-GyfA@#V#^j*d@Ja;!ACff;)VGkV!Xl1F7% zHaBZ_2VEz%Czdl=-m=FlCu>*zoBrNwm!lnP!Jq}YzQAyfmTVYuP(Hi`Qxf~qw_jZ| zNX*rE+cjyYI7%a%r6{p66P&7R2fLPO@Z%9X1@pSrHV7CK3e16ywLz2tmEgY3T&FgO zR`fJrPkoPlL2NhrUbR?E>XspwYmON!YM?j!3z*^51MMoS@AY7ypy&ptkyQuB4K~SV zp5vBiit=OR&ugZeCgo(RrPEBaK{PYzs;{H7Xw;(`mXbfOqvKPGsYZxf$2MYc#RExN zs}syK8!;~bH^oK_1vZHSU#khkn;(i@M`N4OfPh3FJuCEI>dC9<$q=|NSgewnrlQt^ zazs^oaGJrHCw3W=k7#Bv;@_XR;hfV0xla_K=)8 z8kYA1VikmWN^*&XU6VHr`9myAP4vT`;rM-4J1E9FfBlPGReotm^R6zXw4ipUBv7XL zq^xp7_@to39Z03Db=m1BUTc?hwzY1*VQE2XhZ0l6-B>dT=2q&eYZQ}yxhy;zL~6&K z_@gAXs?}#g7IZBndi8(DBsnRCf}W=HHi+1cJ@ZFNZu@N}3)*&lXj^2Rr`u|!W)L*V z5z=~y&BmRXoqpzd>m_Iw^zER1To$8s<|Uw-F$i%6UV>(zf#vh92eBP` zSr26Zsrm4kn+@-5F39E>{2yvslQ9HS9z&XUf=L|QX(`1i)rq0DiiPH3Jg=4i9n)Dc zws}}NZ~*0?W8GK!R8v3(l>MMWKI6PC@yVE8BU$j3i|j(z<(kPWH7T5OnrCQ}EIsQ% zUg;T~c>anlM05=_?Sj~TSg~a=qanxH8&qt^op}C=?YJV)g0_`WA=hM(eVlhNq+0vh zmL7`1H>lW-JMsJ#+i|6&1+iV9VmlcHZD6sTe&YEnw#Wo6I0^l>Dp`z%3o;*SHUn$` z1CKz~w#?*&rF^~(;@bisYLGRMPdtCcc3i<~LEElBZA)tLa?VMXo(&?l<4!z(#dcgV zYe8(+nb=NFs>?Zi#bo#8`V`w4op}C=?K&uQO`mvQ+l~~3>AfP`?2fh1nI?CRo91LN z8r8t>n12<5bx^umchUy!ydZ5uU3ufAjhU>S&O2$7J6=YT`)gRlUw6`mB*KEUjY!)x z%SxkCFK1cjpR~=T0Ui!P)Pl5)NZT|^N)JIX%|B^_@gmX1+|X@sA{V4>MB1k5HcC04 zE$O(z&1_jxuyZX)+c1mR7{3ibnDf_=&Y;{fiZM8j-E9bva1;xRV3PNB=PclsG5zeO zoWvMp-i9oYl9{YR3=b^>M75vNm-8LN+3eXFFJQ(M8G4#w)rfg%z9?9U&fBgaYR!)5nEkG8t-hH_q zUVm~tIfqu?xLV1aTYX?n>=K(Pl3nUBMP?WY^P{vH3C%}uF)pmdL1!!JR58nt_XFlr zPT5a9_ZQ;i${bpUN^Y`Ee~vH2an+wSbdYD>li#KG8Dh3v+jRRvPyQIm95>2DQ{&9U z{FqIa**nK%3RddQBtUU2E$NRR6Aw+y#P>wl1=?W3oYDQyUR_h(bHUc3GuPvpCg&`b z#UiMqCmp>8FA6-smJ>W-QX9KVv(I!!zGQK!-5(cH64;TJ;BFH_DKq_-1S5G#pG%tj za6l)Ge|NtVz!OavpUIbs{+M>}8%g9#HMnBnsOw!%m*@$*bwiTwDM1a?`)cxTnq@IL zdO3rmE@`IXc>#=!++QWn8@Onf*PG}-eW&nQ0WV9GnL>~!+?xT2K|xKGt@Jtv+R zgH?^U_0%mYhcoehR}(T0RRCvq9GxOf5^^XnJFVV+#P&8Rq&K+LO-BDxFeOZ)}|4?OXUi-4#)gizCuX3imeRcbexFu;3=5kmUc;_ zmJ&iU>_3H?YzDM2L@tI{_6GcWvli1(r1z^ItMtVK98>4to6Lt1Jn_szcs3hooa%iETkT)>1k&7!Ht=AF}Eg zGB)#1cR&(|Ff8bfF0pQRt<0pwdA#lb!P2CcA&ZA{GBG{>#A93yYe9LerShNv@6lUs z1KF6=Ci71=W>RwcFXC=ik6HWU>2+i$VU9Zvm3G9>pNHsJ)hU_LefvG?MP93 zbX~+|s_E_-Gaf#7ez<2cfBZ<~K2kkR^9hp8ovcCxgklOEFA7(xiU0SNi-etXsIsv7 za>Vq`{$82Ef$9gc!#z??QZJ$vfpTRD!fAyz0e*BoK^i^xL?vRcp(;N~4C7;_S1XyK zy4EiV=c+qyK>OM)-;hh11)2^*ZabQurkOz-+mIW=Qgg56#$Im4HLZ%bVx(z5Q*j4% zb#V!Du-!+=7wSEU#HSHaoncLDs3S`j+Zhzh)^r{8D`%<+1nYPWgA&;+iYqyFpFw7p zOKS?QpHJ>vSQ;`org9>e(1t@?;7iHA^sJ=E$BO2-?D#woO)Q`V(HzArRiiIamP9$f z2Do;4(~#aAC!8CiPqE88E0)W!cgEeO}T2gp*mXj&;a18C!{WdA^sLP^~JMx9R z6}=yPv9DdlRIOf+h^g;4jta@HBS4$f?Yz$`)_J|JO{r74MllWlU>t^0H^Du5F|;Bh z$_ME$nF^ojpRT`w?9o?JIHrv1o~a8-cpbzKk%0rHa=9Q0q#tEyW1?Itp$T?m8!w7w z^ym)nmM$0i_cC-D{sdnby%utDVIhe|-|mk&xkgI1Tl)Pj$2>pJHvx6%W@a<*Ep#_EBtzF3*6o2t>K zwVlk{L zdjB5ZGQmt03#Py7=ABsdWp^YWx1cSiggG-63!3Q{uVif@$87UXEM`z;UCB5o0K^YtE!^6g`k{9qmnVizxx7tG6)=F&7>(G++4q_Fp_9hN zQx-@)LDA|3?W(?D^fd@Om49JK;63nB|$wfes##GK*iTVK^OHWM(u*&{| z7z#w?#L;jdDwiZ=rN2QT!Nxy|akYk76h*zmw;+jQRQ32IHbwPc-&Dms|KxC7abvdR zP#1MO7Nl^-QpnJwIrDj}4ae0|W=jg4Wqo-;3TG^ZG&Km#SD_6xeYatbTjg=Zo7s{> z9WNHpkhU`auyYu%^8R(Ay-os?KVfijX3n_^}ca=2rh;pydxFxc)tlF34aBAte_Ie_O}PFKP%DAK z?Op1eA?};`Y~lDMCh;L>(z7!@i>BoI$)~>W%Bf_ZWZtFGJ8_S60( z=9(fFsbnVQ@7Gp1bDO7{;ronR)(*RtK)B6mPOczA+A4V-4&6Xu1mwfD5hBf>^}=)d<0 zptlL)@=B-My8Eq#X4){Fib^Fn!Kl>=^s-Vk&Bz__ec`46BBlf)0j**gbi!!VQ$qE1 z@9mQ7Q7Bl=;7{+d%h#h|COB^Jokla|dgg5qNB4_3li4$RLoCLjrA0)1Xk~%!VO}gO zBpl0iu@5#fe!P_Kn)W%o0{i)BS;jZOB4d>VV1j7>{z;+ECa9rxC4)uE=4!1MRxdVT z^jc8_VEts0uZ!!Aoa@Y-w&4h8Tx`RRb_POi_3m!A!?w zQJI5Idf&m>z<$ApanhCWAxB&Zjr?(LyXfGK;l)g9Rjz!m2XfWgyC~@(dtn z2EIJQjaqnyjIYVAJi7rrd2%&V%(5Ih%c^*-y$TtzMy4rS^($Eqrnhg3W`9`@!sXS* z&MYgh!zc-ZCo~j^b-OPAykD4e+rwwzZgT)&S1D9#6$hA#5!ZXeFHCYdG7qfm|Eobl z16&9BPyZ4>^g@-j;d*ud8FsWNM{D{Hz?!-L^0iKzdGIT z@Pd&ig!?OkkyA3mKI8ldFJ@3j9nIseCN`X&F6}LcS?If0kLu2WhcWb$lmz67L{r0_ zy$tCfE40}U0_8b?kupuMCbv?MGEz5LvJL^ogbp$TgsO_3K1$N77ozZz>d^}h%7rt5 zj2cS#_jwu(1CmBLveg!Z49Tw-g=CT-=#9-#jLPY%k#mVFG;4x61xjXmYGMFNI3kvP zg))c4vcGudwpb1rstsL+>IK=v8YX|Igq5Zti8_nZCrW>b=Rn%14XWcD+<*G*jU@|!dNAeV3X~Ry^^#h6xk4a2T z1(0Ag^A$is&V9@cL4!Cg2pVa%>PP?6W}>?Vr)f_6{J<(<#^tt0Oz?)(X^0>eL~T`~ zMyuwlPLt!+5uG;0QY{RlLCdWE@0f2MxB1v`75xit-Wl(#1AOi7mPWHtOhy8sa~Tz zm_7e2`2zGCT$z48K{l?#-Lk75wHeL_jSHoxIel`MfPAX;5K6E1oby*Hj%(3h!w6?O zQjlVO;TFjSKb>Ymvj9OP`x>$irKj|^YX;_g{wYzs-f1bGDc*@h`^>M#2ata1j%peoEuhq^k&mKS#xxT3oVm*bsmS|uz04-8> zlh<>1*|O{er$-_X}ViIn1A zR@r~_WTaBe#P_6?kUBj+;K6tQvlq18@H>znCw3cL`YetMk@w$6Zk$~>R%%KtL^Jfb zR^0hf9O|j^DFtuhu1~HT`lB=;#P)T6KMT#sU~I$D2^Iu-0>jaX+;JtFj5!dU%%G~b z5XP=GP8X{jIM#B(LNUlA4+~4*bPo&591LaqS2?_>CxLSflTf*{45vFC%;CQl`<4($ zt0%E<*}iAh7e<<6CIPd6eJkT>$@*#CyR-v9et9_%9ZA%re9HM7caE#xkvkh+Egzj3 zS$2{lToxu7LTFg8%?~3ugM!%X7{L!Q4LNt<~5Ke^}^gN9Z8hI5P*6CA49y&Pk-o3U4; zFLZ`E$3b$x_shY@$zBPM4|Ht=GcwD4708zLr z&T}@)PqC^AwNj<#SVqOsoJLNf_ZJ4EB9tsF%W(<~tKd{JV&ZySoGiDhq(8$@AI+pc zH*Ub|!{^{vQYoOhDV2gXg|(2=+B$lHYs@TTR1caZ)dO?VUq{F^^vuhUZzn72;iQ+0 zyd{}D=UQ*ytGtN`UkoS_#tKfteKi5q!w>rJ5$=We*{wU))d zC0)-9uGpYfHaCDoX17+xs`_OL*un(CpNAh`^`70R8K(XE-1?$1c2OR&2WPXPgbl5k z3&J*4voSI#W?$F_^g{Da*vPOuQ5fqu{#i@ShMAECVVf#!G(9z{VP!nM`6q7cr_`~Q z;?{-jSTLlcK{U(s4QF4s(Q>M1{s|kD$=uOx)C8Q!1z{T%Hj0_rkHM(>7HKYvTPIVa zvyr&Y+;RH7T+kD?HjFY=%4C4LA*P+zs^Pm!3ro5w@bE|ZoT&-LWDCcdG(z=O+FDLb z&e;|Tc3RhXKM<6QwClb@iDjt?jYdh*c-=pJgDFQfD)j`X9O{_r2R-uj%%L9gE}no` zE`xG=+7L+}Yrrh>m8`RzOch4<>3*9O_>9OG`(N~bkQNA+AURz5!u!ASv5y_}F{?ov zXnp+IGdKC_nZ~Q$f7Jg*!Tmx}|MbrmTQ2(Ar~YxtAqRD~T46k>@Js2lpD;Xm{^8e! zZ~ofRpZ*f}f1faJX*l(;jeh>dbI;xN!7FE6cG;0p?hGe)^Rv%8=&!uuJo~AiI3IcH zh)3IdgnSoovC&2!Kjy0k9kkOxowxYrOWAjBzUip{ znY_KXcy0S0C`wx7@E@_`<)>C|~J5`t3WPf9sd$ zo_wk5`nUdaU47@Rx-Z4F%SPJ7e%l@R+Qm-|y-O3j(Mzr9&G}NP@y>qlJbT%5f7fqb zf9sZ8OxOJ9{9kw8xc;WizOrrX@ty?SkTva;J<Tz24l&$zF9>(oznCFL<06cl)08n%C|*`1ou8y;Eqj&)ujX3**Hdb$B!%=x;b)7`h|;bgxB< z(=NQ=fLd~@saWsmXK4`gQX;uAl$ z=yw-B^UB#DT>FK_eLPov?BGmk8j3%2|CwVoZNziUe;*?%~3#g6YiZGQg3 z#}1je_nOjQ9jEMg;3JQ`^T->o|KRKW*LA_-Me+S&2flUj3!{?rkl&$x{6uEs{q&n^ z|C+vRm+iiN@m1YDF5Y?3jr-rSy>8FDM%O}`tcCmkXY>938FC)=(-)6F@ZI0*m$1GfLn*+=ep(2a?a$N8_5KlbxCZrkM))^}s+n&1WUS%KQJeynM)ap1A6j%r|#E@33z_bK{Q=-iy4O|F@$f5*j(0 z)A-XXuDJUsi0gxUo%fUY!w)}vW#wPGrw;#hFtW z-~NqdpO{o=iu1R86Jl9d`sR(>Jp1)+4vXxw*+cmw-+$|=O)hw;v*o8gd+ukCIq19Q zNB7wA&h2k}@r25&%bw?N9lC_Vt{c67?enI_JEy45KJe(}haUCTlULq**bz~^{#c?O zp{Wnp=6a}#ulQH|<%-*O3Y8{yy-9uWr;fSkkk0X6SoT-`*s~_i-dkN;@!4On^@%;VyXn9kzxmROA8dNlW;RLLfX+)@AS<(J?0@{@nM;OeU%_=V&2__qu1ThGIQf$DvK0sa&&k_-iTv`lUH<-tr+ zy6)i3cdV^Aen)PX&31eCs`EZq{oQdtrFNJ&I}pkA<>}gAEPnUNOOF20ReLXe%iG!Q zPyW9hdhwIGA9$B<|N3c%ymi&jkBWcs|I9yr^UB}7>Um`UTRP@F-}=v4-+g4e15e(* zaOcsRK4IB4ebFHgZxyZlEqUFi4{pA^{101(ce_sgD|7j>vp(_t_;*g*bH$3x$!8Rp zTiu5>Prl&Qz0PkxZ2RGT*L<11@sRtE<60lQ_r`Be+dKUFJx1o)XMVrQj?!jVyAF8k ziR|6GKY8pet?wQF=`-Iyc5z|iq?g{l@~gXl?z`90;_mJ3AO7A$ilbNTe*3xK*>)3r zv-}S~+pn+KO`qN9)@5&g{=7q(E!9VVp4oaE?bg-X_K*(P>YV?abGY|YpFHu~kNknZ z@{f03xM%#36Mp#0*A1KhpEtj5;%SO5LYVSy8G{gFQLL*I4vGv5BV;*`z5 z9=`N9oBijnTOVq<`Q_!OKX}vLn^&*+_F2|tsZ%58JapB5-?(M(*RM5rGROP>wi9*Q zR_`A2R^;^KKJDAYbk6bLTiSV5Xxx6sdDmR|uUr4E|Jc_axclzZf$#on^Ly-`I8YO}_ikHK%WR)+cXn z9e4B%{1JVzRah1se&NZS?)9C1+j)NH>d9YxeT%~^U%vUym-z3PvpXNX(;xqCzV{^V zi{GgJaa)xx@F(e&&(gPle!Bzeh4!Gs9BXnP0y601YfYf?>yXU-rJ8RcH zZZH;GNBzGy_Acs6PcN1vDUwK>d-uKfZmzzs{nVLz_;20+xl3-{Vf%{@{pF@Gn%_A2 zpq*~l-}>SepWk_l$M5~$hZT*!-oUVKJ zz54{#_l3QGz3kz|CvHUjV%a`#9T$1$z}-%xk37Zq1I-_Qv(3wqr*}T%@`>-6&%fRL znX~tL<>Qxb_Zja=)xZlYb`R=r`O@)WaYbuMW78`$j!Rw=_vZH6FVbQSyIVf{kn8j5 zKkV=s>f(!kd)yIP&+|`bxhGF}{u2i2+EX7bEPnU(-(G*k8`nOy|Hjt$AKT{or!N#9 z+s(G`nOEMl`GF7Z^oKppl)ih$DPK9otZzkL|JfanZvOg1Z`}X*#mis*{axmFOs_1z z$i2z=uN2?kN{|k{=%|;jd};4oHe1>3QEvHDw>`I|<>h*sy{@{=@Q zvwm{-drx)!`C{XizpCkfxW_m1=f9jhN*8|SHgUgm&iwkOyIlOg=f)0>hR4euJad_PGGn!lXwiFvNy zD3-j!eZsEHQOi$#;r{Qv{mZlOq__FgEA(cMZ@u&VKh7+RRUSC$mdXtu-{d~x(MCVd zKP>)f$@5?Og?*pv?)v6)&pmhYuI?Wxe!uLn9hUy}i;o`p%3t2E8qRy+s5h=T;*IB* zZhY0F?qx@x`>RcU|HGG_edWnVza4&H^T1Q@7#=)z55tN-?sn`4)<<9a#68=5=8h}x zIP!AmqwgO6pz)~_{{6Xq-aZnp*#i^TJo}I4k6(P~ScuWVw}v4^7d&2D{A&_q5Z!yh z#vebccAuikU^Ia@hyf9>hxc0Tpu8^oR=t1H0GX|K0VTuR2VN4l~~(M0U9L`iUp$^L~}yZBg2E_>!$&iR}Br zapBtv^y!NCmZ<*k;3bxK9<-lt{Kn2(KK-}HuhpKqtg+u-kNxJATRvI*;`95Skp1<; zKcW)9`Sk@4i}(NLThxm8bl>^+^Sd47edP!DdLKFPYs^Oc;)`~=_|c=@xqpuX-?F}Q z$0x*>z-+BG6^d~=e+OGe(ZFj-3 z+47$pe*W9+=>M)*@m;UuTIJF&o_2ZQw5W0Sb35M&Z~5{k zAK&A?KcCKA@auCQ{K1zG%l$fh(ernIahp%kFWqzXR^n#2{Osxf{PD!ke*d16zOc*J zK7Uy?ack7L?CZ`StAD@se((L{g4or6{?wrl|6s+ljTg2(;p@k>F8%Y}#&dqZ)4A_& zwY%$btN*-L4}JOKqnH2d8{0p0-3{Acw&mro!Rf#9@sqz2`_gHb-swK@;pH!%c2VKa zTfBGWGMMDTMc2J9?tbU-J9K`z%O6kt;MvRm_5kGbE7v^pi!)BV<|MQI#q0Ne_ptB2 zasAoLuia;}I~vPRdO?5i(tp2WTyf;--BrD@?cCa1_1Eq_>Zk2D&-%>nzuSg?{?P;9 z7goHpkGp-`)7!oKfMwYcOSvDtr90#OEf?>4)zOC>_UnHfbmG#zlQ;cN*4w|?qT)Iq zzSlZlswokd6^IB`DoJkQCU+(Iqv~(E~<_bV~~eJoEdx-_IYgAGYh- zIp>wyuQgNY-1N)FqPhV#_RY+M}J+a4C3sN{0b^ zbvEr9w^T;hjOHT)pB~5ZR!v}@zaR#TUq&3n+f1zSX;eKF{z#k%0~0er2#~#ZS=>KP zuSx`XNjqkF`O{w=8h+2Ix#l7RFUIbmC()d6x273=12Ywkri(XTgn8^=rSm4BB9?OY zSrxI~8?EpX&p53Ghq?pAUL$&W(jmq&a6X;K)Mm)s6c@t-_i@VE+uMU#Vn+N(Ba{uP zWP;nzhOipUHM~+Gpc*ND$3u%9W z7mqx2ln|p^IqHlgKmzsBQ$0?E<<%2jW*}pkj*Rm zMmziZ-F0xYZJK=wbGc3Xyu5apa}JsAcU{9#4Mx75uFH0`rKWLJoyFR)zUg~YRc|i( zdVPz?xO@pUQ!_xndO4jS|H%RxZSzTfk=uC>)>x|RtZfFXLse9y81ADm>s)%U0L_vG z1v8Z#xQmJNr{Q9QnHgj411%I(@Z(&&hEeR}qqStnE0rWom^bIy2Wj~VM(iVOz#I+)Vq9VJ)Zps!f zjX%HpIHaw$w6qTCPSP{ooO63qTF&8d*XK&tzgwN4f;z%Q8>u8W&;HyHvn6+0AYfg= zqcdX7A!(+if+O{AI#AJx^BWO8WeLkWVKFf?LlySlt_q)vaI;0v zj`s@rBECmqS3nNY$hK-ImC6L()qg;fssZmC+uQBJ=RQ68FLUSXKZ}o5%KwJzs@B`6 zpJzTQOJUBK{Nun4!$P`>_mxWAmEf7-4fITw!D%yUVUve#(matdJN;euu$pIufP*HL zTiNidUp|wg6m4kJ65*OG`ItfbQKUu+(6=ypGIc21C2XDqVp8?So}Okca%1FR!qKWR zM5b%dP}pZ~vTBRRXwu)BNe+A*^PUp?ZcW7j+%=(_nk$ve=A|ZY^DCeP-h9>W0pIREef?MkF+AwyKzZ z>pe};kp_iex68qw)3Kwam!&&?5!fR;=`Nj-k_etva@1l#*$w+j>s&`~`vns@5)(_-e+>w+R)8c6#X#)1{0cPj$$gtTG z5a_{~m#j-rkJ7<=uk+{{+i^EHl|iYkrJrMGO_)=XOIz*oK2eZ>R)6f5b+cOQm5mb5 zkIZF@#HFKukk5#qQ^CoNlGb22J?Z!Y&+HPod{=$MXWgsHcrX0~ugn~PSTdbeeeM^k zEV0$QK9oW)_j^1_+O7+W_5+VAb#q|NG*FAQhaZ`78be&9cAWsQeVAbT!F8sVUBHZaoT+jOlP5z ztQno{=1kWOf<)(M-#0uOZ!Z?wq}-K*R(m2Fy>@h)2(JZ+))b0&7<$G@jGKSkL;ehf zEYK1`PJ#E0iYhOkt3lgFG&OebKN4HU$;xibWJV%CcLUcXdvu z_QWC=Tq%%EiPUBVvm-?uH3D}8f$0BBC(mfXn%pz-_H@inxj|Z<%+bsRS?}@r)$ke9 zL^*_dO=-+rV5)YdhNJos?p;g1Gc7+1&CVuuD`&D++&#=^mZ7?jl@R*mrcj=T51yb>@-?WCS+G-174tHDdKmXX_?4yxvzhT))K{?6Zp)(&G!Q-M$@m4f1JVC>O|)Z5b7 z02#4U^vuk>BB0qeSjROhq1NAl;LMsH5#FpfW;*}M_E;ie$SVy`%G3&@yR}fKj2K&R zsgLUo^*+h7v3REdi56Y^ShIuoOYkSJTTn{FZu5v^^`G{lH^HSicJGH_M7E5`PuIcn6Kiat;PNjS^ zVyWMn(Xg!BERATWsv5A!5Yie|V}vCEyqhMIF-C)+G!K-35mQprkPyc_E2u<+xlF^# z(#cj>!?g7K6-Ax`Ig@raqDfGMj{;~Pz*|=#oW!59ZYNi4yr*X0#*;+8Gg@2KtToJ^ zL3~rw&>sRS8GgTDBEPRKIE5O`eHQ}DbQ>E2q3b@gJ69#WG{-EN8!2a7WQmPDZ?iE^ zLlKXq65kbm|GI=?k6kj^${`a4Xh)iO&EffE!;chcVe+cW-*=J{YJ&^b`A%k}H}mi~tQy4WHj6)N-YiH9J-Y z&I}0re9Fj@AbAoUw5ylz!T!~E*?$WNi5SfFMw|HDp3k;?Q^587Sa75EylU`+EzVH8YN%$iGK+Vol-+oW+UsyH`4>#J zi7I%E7%SvQ+yGv#+bvX{^8iCg`_$q|4-T07pqxdXl|w6gJ$fBWB<=4qrvI`0OBnhS zb8KXdQTC|G)#}pgyoDK#ln~Shk072P`i&IFTQZ-#Pqwy*4t^SRuBY zWy~*_-mmt_v+Ein)z`MCq{wRgEjQ^S3hH1NVVt<{&Txxhdb|apS7xV&f?qY zjBu*ml~`tf9GAhewwDPSoCN<}m~TJIDwlRPNA_eHL`HO#3zcFJfT`Tge{h=hH^-d~ zVaZ4X_Gw@gZOX44aFQ}0rzuUV70qcPijQn>#vZ)BC6B6z)n=H?DwCdUHb-dEd3lp> ztgD-*8rEwr+>@9w+1MhbUN@@%%j3Zs3RdPz?%=w@5JIE^DrOEUl@ zj!B2@l9uF(@g;!<5GDL6P?p3j(*C6VL(8F!I!11WXd#%u)5U#>@1i-jE#KGzF>~+; zztuaPeCWVwLNblarBGP9)=!ycFO1jhcdTZoRikkvip+2b7n9@#%G@h8IT$tZz!*4I z&Niz4K1fV&TSCSswKU+KUsnY9_LB<+sGI2|sJ#dKtGs9}39@Xps}-qE#}_j#FQ4*S zZnn4-g+od`SVImve?{`o9W<>R9&!esO{g6FB?e992A{^a{1y0?5AM&(WDdR)v4>MT zul&kc`Gfp^em13Z3HKNyW<~ug537x*A_zU=VvTzti1&p>`JatxGR9P5X=xU`AGrWe zy446eLBf1$u>Wzbx#K2N0r*M((SXxgG#Y558G`Q5vccmyZ~tKZl9`pF5Cx$fx$!{N!^2#w#VSzb+e?scW=4k+Lgp0If~GV6%J_enn#bxj?! zs5ra{Tk0O`+2OtgtR~xXo5+|I(Sg}Y(%_Q@>gBo7Zd1&842LzjcE_Jl4oMm$wt2ki zpAB^@t_LM0_yoF&-NzpS3s9jH*VjDmVjst=_4Y`HhEeQDLel?6|wnN*rw zH)~Y{Ec)aFzEfzIaK3&ot=;D+iNRHK905)-d(;V?#TQukT-ZXc`0C*usS zQlOrY<*bBOy*pXiyl|eLWwTeUiFaPKTZdMT#DMw%ug6O?=vjWqWZic##Z*ivs-dzt z)YNv3!^o_+J*)A9yIg#4o#CrMZuh&AesW+jTlEf&$V%f;&=Eu%;&(n{H1a$|jPhpT z&tIKs@O_9kt=`2N!~NXr??G6d;^yEZJ66E>Q)c1bPu`Ke8%!HtAU{aMN;zdc$XiTU zhZoA?m#n59(I>A7c%@NHbl~5jKu@EO$3$bg%ao(H-a#*@W_B(3#4by4Pz1~fgBDE2 z3*bTDB`c3muQ-Pi0oI9C8YlV+mDmfflw$U@%A`o_T{Vl!*XRU_(;H4TX{)ayUkFJ@ zQ1y%~gPwgUrH{z)F7hswfqYqhf|-iHOSD10OW1!nV3BxEEK2{-Qufa!gs;)Ls%vuD zugBWw)<6vN+kI`8x<=#vvCQ4{+C`Mi?J~)lH+OGu!Cxiu&Fe_z|B(`tB+>nB(du6{ zqVd>SN8yu8ZD6fQyckyYSCuZXh(y2?+q>anaf;bn=;Y*ZNm@Va8}?%ID&nkuJ1T22 zzAdw%&&{K0-cV!Q8qx%#-IT;No^}Oav$BT0-4Yn;xvv^a?#QsB_SUD8@dn*0y-%uD zNU{xY8Y4W9x0{RpL#S3R^#=2pzIj=|qB}(@_jG^_@JZ}Nb3tcov{GenXgbt@UHknK z07RbCx<5X@V5TuZ>b##`RJE_@tTk2>uV02zapcv5H6jpOpSW>@;jBz?^UWX+zQ*u~ zZ5Z6VqYv~6ISrC&DQswO9tRp#jb}Ll(32)MvA34#4$WGq4)ssD8@i^KHXuw&3OL6E z=)7ckVbi9+-fFOkwrHr%ysOxYCYm+(jQgv|qGE)z6+8K*3mz&$ZgzA+FxLAcVj@Z- zQR(hMp{IjZH<;P|*1f9w49kF_EZgep168i5f-h`QxB)g#@gG5SRT&-@J~4@WU04#S zKbaw&sJ8pMq*Tg3y6|XVj#zb=F^bq|keS3o4RYU-a)~RRE$C|Hu|GQ^N$e5Lx$brM z=gO7dJ|FMe-l!nH1#tf?NyT~e*+V?wjpM<7)~z-82flP)`h(D@7dKJ`1S)2l*yCq! zt^dJIC~iH`CTsE*M1>xVPu%wRouPB1_y4)t@lX4~*&5`Oss8HkdvR(_t^9w^iHD)q zv9g}e9D!^<*tx-1VU)$ZhZH$4-cq#j>}Y;4)Yh>{BYg}g%1McGGVoGW$ocxUE;HJl zor@r8&{`d{s`&_Nmb3m0dDH6H%tqOLsv86VTroryG9ZVQCxorM}iL1^>JQQP5p5MU|gxfO$v z*id7B2b?aH9M%E)#A{wV`o2O5&uf`?cG8s^e^%u{c%@VgvdHH)4d%_efVCUn-;IUT z8p2Sv1cTR~<}{7$HUd3cR&;dh~YF^tKk9Z~UqMv#2A$v@}&55YnaRE7Jq2O|B%PC8ER zRgbz$i1^A0u-X?}y_6zH{ zk1Sl`G(V+G;tY6znL9}%Y`G0aciTTSDwQVLl&~eK36C!cdTJVt$Ytaj^((<>z2n8w zIOnE8Q{S`m)3`jE3^tSm>svpw!9Dv;<#aVE!*Y{lckb%bvqaLToI*w7RQbas7T$6j z2zr#INR{hxW!ny&F<%1tzXo41JM)?%nPtFK@GYkzT!8HlJw;8B1rc-bg}7sqWM;! zM47DVsPhJxFp=4H_jkj%K%&;&^}$5sYxxt##6+<{ z-gP4Rtxxm^jGef{OyR`3g9SQF`HILe)u|`RcoPwfHy(JUL{QG_)D|h~3wy*lzdGk0 zZFw03n6Utsbws@)Fxootif+fs$IGi*H0P?hFQQb#+`T}dxe7zx0SOhKxPr;g9haCNsS zIX+l(_O?$R6<2yGq2!P;xQad$YHHA2kgt4EPM4C3kGONx{)8J7lM~Mw!}v;F)gY-& zW1d$wCOYliu)U~x^1G)b?5zQ$8+km#KGQt>%y5O?bUkbBb+MHA`4>rRRga#57jp?A zOFkfoX?*)jr{t~qpT;@D!SUa-Bv$IO%d|#YJ%)LbHXJIAU!A&TX^Tx8&Ra{kI*#1m z`a-U1Lrj{YmDG(?4fOTRAP2@VvpR_Rb2685;T~`1KKx%vjW9+PoAv%_@KsNL`dqM= zQ}S`m=%JCnOtB+p=s%h+62HPq7zWK@r;6Jj>j>( z$kr9^Kggy?%8X~C$nunnU3Q%@HBl0ev5!Pvg+b>f=*dLn3x$_lQIgjl!;&NpgsNP& zB-9)wv1i<_lS@4 zp0f*GomI<><03PwnwJ4`GzC~g$ir{=2LHP+0*Jp(HK_aLDi;mSV~jP~`|-2xcRznY zpW6j?J82e(5|G~F$rJv=f*nbXj~2k6a>pO%%CgrO z;?V=)kL8L{)Ou~jtRk;h76Ct#FUaTH7<9wfiB|Ex7Fs-*pF@_ktY zN^{S&(&YT#L%rY$8l9*uX3is+pBURA!x0!TWhSVs?)8dq+X1*t?fhb# z|2`icxT?07?(2%@X6Y-p#Y6{g?JNhCvmuA^-zveuREN*02xhBZmHBY@+8{7+f z#?&RCXHG;dIMxMcQ4$K~>Zb)WBZQUd(W%Vh{Xd%ONStdH;7PgfIQ z8}Re+l6*2wrJ$-2<$AZzd$HAZH}o>(L||dh8?vwrIkeSSuU@^s{i6+;T2;YZEF8W& zYWvY87W`|x7wfQe|IPe=qAr%Olg0CD^fpzdvDHkl&V59ptQqu-dDuj_8wAQ`{WUYa ztIk~R{uU+bep!Qn! zV$bX5iLL*XM|RL+kR}?md?)Ie(F^nLm%mx#znjHS$~gP0){K6$HC1%1vs6ns4_%jonKv zd`Mc0CFK?*R<5)iN$KGWbAPbpZ4d;@^dBJt#OHVUWq&MRc3#dCh5S}nAT~LwzW>0Z zFoh3b^|Qgle+<#Bo>!WB=6>!l4V9HwVr(cay`Z-nZkH|H zd|mrbRbmGc@(VH=sU_~#o94_CLfV3U_8JY3MK0G63Vb-fvhXDU6I*<79iS6J2sGLq z$;LLZVyvtFBYEfElc;W26~pM>vMFk|o#u)9{GqCZC5OY~DKuhU`yxb4+L_q|7aUT& zHmmC6gz5zFgG?%6m38i9yRdo_(=3BHbe#>H+_bGoJsa>mhCga0-{qr;wjhIk<;mmAk&eY zh_1G8UQgC|O!xBVX_ars2w&%Kn=*pFn-f(cCl#rOk>5x<*ym)XK0feg8#A#}Wm!)b z#f=u`r-%-W0%_3VajcK$PerE6KcB$J8&m%?Ih3o2o)kOp6U`Rn=$oq+Fjaxd())KQ z%ZS<&>yP#U6f~vw)>m-w_>+R_3{hmZ*Q41?C%JD`%>Ml9C2BkEh2;nD)0+P#Ow!#v zw&7v6cbaBeV0um;s$+3^Y^h5}Pe2;jZ9<93&GOv;LzD~8Ap;&D_d2P-q$lBujKphl zION&?SrNYtDi`Rd$*=56z8QE^xm(~Ch5MY&jg76{2D%8Z@9{qA%%$8}TXc(xP>kpn zm}K@H)C@Q{5kVtt(yI89mDMikC`s%}z{)w<(|!UV)dh@^Sx{@QZe=CBvuf&MZVAT3 zQf4qIXqJKE9+vRd#H=O@Pdp;Av#{oAb0jf2HjLk7&afrV>v?JVu-^&VhVs!XO!+quS%vE9Rs3&Re) z@iZ+Z{86O{t1NpCkr3@Zdbhbu7u&M9+(X2Ypu#|hAzl5+g)*&?klX=VU;@{CteW_l z+Cpjt^ilmb3EgWQ>_E7Ge}mQcD3&(NR;~?wx8S}M8yryx+3|4~c}!LNiMj+JZFpQBOgN z{D+|{QZ*d9V~)fGqpJR!AXm&~Vq=(V@)Xgj^+#n#btv+Gx*p`33ZA{{v4_{kq8rH2Saap-lu1 zZZzySE<@3No=#_s{$Ld03VQoYcLCW?k>{i8OPP>ybsVZk))>-pNa&iOtd#xp5Qd`H z_#G=@1FK5@L^jx;`19qJQ9GLJb6#oYad|;ilvN?U5130&bJ&!OAI-)mM6oj9Na+}( zL1SFlsM@3oj}}vR0GmQ9(fmoPCA9}dbH-Imy{ID5#vGk}O)_00y1Q3eXQ z_tdFB6Gc@ilkK7viSMT=@AcO4bIcO+M%)N^GnHFw0+iN3Z@ARC=+aQ-1zOf2ONjrL z*G+7TBD|OP`AnS1SsOZ9P_;}FN0`Kt zfTZl%Qe)|h_mJint}>6=SaA2DA|+5Oevb2$WYg2`c}NiHsch>+jr`Xu-+xtqNIY0& zt$W->s1tbW+gEmGExCCT+U};4Rc*_o`pQ#K49p0Jr3uWD*=2b|OF&$)?l|Y5!u55Im3+XrNP%wDQOt2>ijPz7CtRq)oENTTpns_L)%rmrf6&6kyt9>AtIJO1 zV~uUpRe;I5@Xmd?3d6k-BeGiFsrM1qE=^T5&iucJ{}5+Lv%#Qr!(3%ugjaN)eYLR2 z!V%~E35V&U+QkatEpN6y^_o$i?Zwuwc?*;w)%y+TJgAd;ouKWY5JrYLTvMqu!(4QWZbD*M`(xX2r3 zD|s&yN?iK5TCWLBRbCmpR=6XkKr8*y8F>8Udrh;8Nls-g+2poIY4%OZQnu1ygYYYF zjj>mPX5-XOgSEA#grF#i7{j8Try_jnElUPjlfobkXX&z42g222=rFlApV_S5ahTkC zGyBr}ofu768&QoqcBQDu0N7ZN#r;xccQvny>fVVWai?Mqne_H(=s)PC8 z4bi0vgy7}Ge@h+Gv)BSR&AWDlOJ0qVU7QqVm`mlW;n5~wb3(KUCfg+cQ4uV}p(Xr$ z?l{hxihR6bx#>IgUhQZmU8B!Kn!1=IM7x%)a$XwcZmwoRn-@})-%SsYCAs%+-Duv3 zJP)%V`*$T0J;LA2is`Jzj>V?%3F=LY3eO94hrN1kh}?V1XeXqHKmHymv#rNEIayY} zE)CA}?nt**$>tl!2zxD@w;8o!TDfbyY9-q5B40V1M%J+p+56k^*s+s~Tbp^x^=UeJlI+1~;sNiKO5xGGN#I_q!-O?lC1kFi ztS3twHS$w%tqRpJm)Y1plimt5U2m2r>~3x>@9`(q>wQbiXhuLlu(!mOjjSJ&ULe7z zb0lDUjmBCgZ4#s|GV{D9)le-Iid(*2H+gwJ7?7)zDpPz5qfF*V{Kp4;i3*+hiKc_B zo4u)WjA~p@Xz^Z#j>QENI|=s@+rD1Aq`|>qTK64Fc@u4iAOd4Q&Y|*|X=D0kkWN}K zo?)acR_#MqWp+3*ukCHng07BT!Zew`!}AMeb%wC9`p3J`PF?d1?5q3XvJkW;L`M2* zTWF5Nf3+t!aJKvTcYrg#^%TiB#Q+rSaoZZ|2m=|r|5o|fwB0VLWAC4>i}a(W%R6lD zhFTJ5)ePtdlOOr-z@zHMdqK=`zH5;J-5=G7e{S>o1$#VdD?i?yPPd; zTN1lNQIt;ukbozmA)F1VjIYrB~w(G+C7w^4i-IW~{VQ;VfqbkkDIj@eb z?$53MfnF)~4wB0_rH=uy**oP&y!47k7YlB_%MsMuSPT$vfsIDFx{TTZ?)|Zv>w=cp zk{!eEZ%_7x@~=kO$HdHm@)*gXaF`sP|HpB;-ye!2xkN%ef>)Sgh;5?_IbstnRGGTV z&K%o2`)a>@>F(HnM|bV7tEmpT7NVSZf-GovE0(0$9>dJXxb1C~(c9xjyS;dK9HJ@`P@fk55Q}@pzZAh$`qm-$Vt`WjXdPZE( zJe4?x5#K~hRaWbh-;qEFM@~xnif$_~{j_Q05Cc|D(XP@`=zah)LKS(O4)>t6sO5c16ubDD+=AS(D;RX^2 z(pmJiY0&C5RDB^;dy}6V?cIiq>DM((DX$Zq5A;?0fUb3^9A}^uCh2hp=A{w|IUe!| zBKz3AJ+-Wb>Tt*E{)_d@2Hj11_bUPFjNnB3d*<%y$myd`M>O5D6X}6?3JO06qUfyi zujgibZYibVo(FK0!ukF3G2HvH)NY*0f@5{7bza@C*M5IPoRWrL?Ei}a#JtJ7(f4N081?Rdd!4UD|Ao_H=CYO#|MG)=tcJ_o*2?J zafW@q@&r^nl1yjf*^>~!cQsyn!D(x0=4pfu>!`SSGt)#w4o!lbbZ9vSKfiEn7=US( zcxGx)D&5Vr58_Tz6?GhTt?U?~p{ZLA- z`>~NXpdsa`UVzIk8j?aJ^XJ#{>2RiF$Pr|v{;#7sozG%-kcdXV$BBuLzw1vOOYM(; z&{)s$qNFwJ*cJ}cM>^6gSoB-@J%Q78qRH(6uC5pAX5p5uANU)N2I!zpLNZCRyT=>{ zv_TR&VL=STK_*su_Pd!00myK~wNwBOR#VtJ_3v^+{whxQ7K#sHADerpD$zxeUq+uK zE77HGL>%lQh~o(LXr#X_wcS^dxeev@JAPz_wP97l%j;VR7UryC!Q8e5 zh3w}Y1(x+b7|{O9<(+y29(7z2QfG|twT%Lv9tEoO}bQ-@fT0791ZkmHyr68uV*X7yKd1{D20&O|9oYh%{3@2wa zY8%qYLt%;RK1KoyOY7dIQ7PU*B%Y}veBvL`uRBP7e%9QN(7=POTSW8ZE`cEDqw^bK z?N)1n&mLx0LeHjVikuuM6TOh`ruHTNQjEYU!pXhe>1qu-%114YEV|Wg+4JEiKO5Kw zroZ|uEpnZ|6>^>q$nlOeh@!}|6+})Fsf>H`NmwNU{0+C_H!@$r8w`-N;JbE^s56DB zo5fz?MSPf(z@t1X!}_-*f3mEE8=ZLZbFcalR~4029+nk=uFs{JLS6T_EBD}8Y)Ck< z=4wmkerJ%{NKwxCL;It7z{w!X!ADxxnxKpM0ch|w=;qM3%h`9ydp2>ChA)q)Jy9gs zjS=!=6`$EIeWi_)>kKy~c>R-8R#7*Z+%Oew-09zl=n1hqULnK4n3CD`+hr?gt5Zk* zf(V)j$?6xjtM)adH2!XMYe?$mWGq#n)QYG)zqg6$5lg4U`b2hp*H@oxNx#?pV7$)J zqqx?I7r{zxeLer{0{FFdG|zB!6Hf2nXZHMbs}XS z09s==L@I%GS)aG;J7J^zIj80aYMO{8)lr3I##r1zzHTykyqD`#lY=HU(FGh|iJevw z1e`R~tuEank;PC+U#hjf7y-YFi7}Dil$;?CGU2Gq*%Vh}-_ zNay|J1S`9n5rs2KnM-n735WdxM)55v$2p-f8L6+0op0aft3|=Gn|rDDqDJ`JY{pNo z1mtNn%X@wQSMY#?L--I;l5nbd3auyaqUQ|x?Cff+_w^B zppsllidh4RL60kH-B@Mgf%Fnq3ITMmKi7kygqSZ)gA?ewR&_H=3)}h65e*|%t&FR{ zqT@BZQ;bTSX6-9~l6Sm~r?CN$CB8KmT+@(1McxSni35Cau`cY`0`Ot8_*98x-jvrYVN9~ z;4|Nj9>bae?4G>zE4an#z}waH(cJ$SNk0qIkwW|Je0)BZEmE05cbYX$n7lm%oo&3B zq0`eb2}e;(5$UQjo==dU{ds3zXqCdzo2oR#ecl9+|MhVz{k|1HZ_`$KlRAW_uB)3DTcFr_0D-x!ECCWN~aYP@mXY9r0$qcse|8471SzB-s+H; z<4p>S>@%w{<+T9+^ z+W?Ex)D#fFT#5s-WiUexX_p3R1@H|y`Y=0ZO0(e}rtWrGp8A zrHZE~(tn&vLTog5^MfEVP^+^17=c0TqZbOjNWJo|wDLEJisU~NWaEX= ziNh}egiJHmKM05VzS8BBz2xen0qC!h5;yRG5xt`6C*uVlcaY}!GjeZpd=&6o| z1-}5;!hoo2iaaEhwPCK_o;P@UCp5ei?m|i%a@N>$_O~$97%U?#ycZ4&1uos zFpbj1$OtxIvD}(@HwW*Y#Z5SQ`aq=g)sN(c5zF( z36-^C_2(xgEdjm*M&|x%Y<8y!+g<>u-|>EW}T)`;S^{q{X&H}B;|FnDG5$Z;z-xKm^lx_8~=i&%KV2T8C9Igh)DC)2S!%KQ;H z<0u2l2-()#1Cw#1|1xiT1MJiP$+X`% znsdVY!{}z;zy7;r3tEre1k(UcvlfpMmUbuA?ig8_=xB~l*d&;5>GW9{ z!5Wz=Xp7}`du^+X;2`L%cy3HV!UnO}RI<1=Yoz0k$YV%IZAh#WJ;~KM-+(36?&$nF z_#-_|x&{iu_I{wAhFuW5UFa}iNdC+nTBqKAotQ|`!ne9VX@ZxwHa6`Q_CLc)Si!Qo znrndi%{dIR;o>stWsO+J>frJnlaTmdfKI3V#o&Jc-3FIcOucp}?i*>?Mzpg@OqEsA zC$Bp7S(2ULqGI7nk;ww>bd_DrySEbHfD)-rk1zC1w{|&SAo2G8yk&a$8onCFF97?A zVdad<&&ngOK113(zZFA;zUn&ac`K2h+eohe{4HuAz`t>)5fNc{wXbm_S zFt#0t_s-VC$P$*@!kh}Qv8ll2V1mxPd}?FzYnrJx0Zo+}V62}^n*?wh&q;ek)Q|-{_R=Ua+C)la^B0=ku0m80(d1F!kr3&Qw*XSTZJ! zIG+VYag9_6)xBPP(UGuLDrHNq|3tuH(WydZi$tQd=fW!z{LRiBft_n0KD_UWUPGj0-WIqgM#Z#YT3O%@)tPZN%@mA@zvq!8)p`F~N? z-LFL-5tUdPeCeW_DELDPQubeIY%7rqw#%S_1M1fUN_;|t$KPXMU=aVDaN4zaT8^}c zN&QWE_~I77uXHThhHDHhbvVbzq1$;`lWAx;C@=$b9-r5R*KRnkyhVUqYamsY*@ts{<#hwG}uU z)w!P$yay5ZCf+OZVqIbNBfnLbS|63BsLG(eoy(ne+uS@Q)FHp)`tbtz?r`t<0@g8M zb-k^2ShyxlqUk~1B0?i zq912EyjS@n48r5Kg7I``IeQ2%s!V5m{BnOc_L&n ze*5DJ1K28z+nso7JCwKKu@03r@xM3Cz^i}plet^X%uEjyM?Oh4^iZ>8(^*poQA>K! z*r4QdQRTj1+hcmb=EhrFJmaHA%K>=3n(GXb1UGWZkEUs(VA(r!85GF8F_?%bPM+K9 zlKi-m$IzXQHBT*)$2$r=Wq-#e@yPB!=bnIZA;9D;41}stFG?Id;vK z$0(!M@{62LJ#~@4iEZ&-cPDHuthv*W-OkrLULGb~KwnHPS0@Pr8N7P|?KMCAvH8z0 z-*cs%iyCV`OpMUZ1Q#H4`K6DiZTFk+fx{%&;Y$OP=Wj^oGqbtBzGOk8*&U;vN`SlS z59((-{~;jXMPoZJkb4i=U;NjQ^=(`J!D_YU^9DEZoT;MpNXOhaZ$Di+A)2ORa7?2R zU*X)&sQ~Y^!R6;!8kBmZcuMK@uaxA3fSzTQS)ZB5`*RVh%%On1sqY_jYLFxrd}M|s zD(enx>?@X&?eD_`N4)!*CCh9(m>Uz#BhDj`oX%06%$?eJ_A?cVxV~E2JjPnwww~%E2Xk4E7CbeZ z*z73t$Sx-#50}PZgV##Yw()d(<|U)$Es&Nza)UvnqE^q#E4|6E%mCI$hmY@m56HOY za&%%V)Z%VFF+KfzQ01BIV#`MCu*tGAsrx^tdoDRgf9u;!%0As-OTyZlYQ|=T3gPqcj#ph2Etmkp@ca2IU=X?SaU?3sKq^W096j!d?Nx zbiFI{a^vMnNJz$YFtLqr-x-SzD??l`Y!&L`vd*n z_OV@*{7I;iGlkKo0a$l*ZRDrixRh;4dnvApjzrtPg(G5&PdV!p{^O0 zZ}VN)@3nxEx&NbO>TEKnXH=_rGJJ*?|NjwnR$*~Q+p=vWNN{%v?i$<^EO>BtcXti$ z1Se>4cWK<+0t9#W#%Y|}IcM*C-gx1u|8K3iW>t;C#4le)MG=9=0@V^vWxn<}lL9XnE2+BgYu9US_#TzxF>oGO{=KU}FD&9|fU>9Av@TOL z;fK8Tas1z{-L;Sm@G~SQXujt3l9c;W*i946;$Hmye|sf9qflL7%*W@XV?<$@mVE%DCurIg5|~-kc4aBKz||^bjfYd?rH63;%GR z2zIxbdtpT)GWyfism0chua~EJYlmewUdjk;r(i$&;7i3E!DTNwaL)L0@Mie5kMcyr z#l=m`*JVev^K-k`#2Mz+n#J?xlGlThfZE)(JCpYJAdyuUR>9LQn}mbcd%hk8pW)*m zY_x3{2-pc>E42E*Kk^-Upi7n%ncrV4Cx{7aCzfWUuCO~!;-C}VAOodfj(WiLf~_wb z?=lNcy@Qp~JvywhlfuZ3;OCb?kn*g$D@~!qd9qUEis~LglCc!*ykv%C3SW}>XoHrO za*<=|C3F7AZ>g~^T8zYLa9UrAiNr((Om`K`suZ-5iWejL(REe3(G9MXo!o(R`Lc8^ z%6xa;EcIJC39&_$xUtQjd(Sy;SiU_oO06u=_r8U)yWOj%2>z9Ms5u?eE}A0b(a>07 zSq^pYdH6__Yv&}+PXlu*2<9;1%)xO-IKCFIiUqn2I@(MoUw@WNp?yv-ri|FV@}Uaz zfUq`CZEbc2BI8{1UICNYg&Z{YM~h@m;B_(uRd7$}8x2|2&FR%CdC`UlXRbjB zE7+}2SW(Zl(lXm1T+!;#AFWAYrOy7uH|K(RkIV*U=+1QU&?`=|n;=zDa0g(l?1{&h4`Mg5mWy2 zW3pwT|8v!?sqy0JItHzK?#A9joS2LM+?XiVNW963>1S_=5YH^C@QG#7*i0p6)@3PY z8}+({+%$G^Z=#L>wJ6?4iYSUU#A@@zn3FwKH^y(-49hC8xM6h874Y?s%DsE%y8YNXKhf(ag_IO$saxy?xSFJWm3l&U>OilW#M1-z5oQfNrd=K4v?D4>lAK!chdHBslC@wd|+j9T9cqXn2et{5`2vA$d zebPsR=}qyf>S$vMzdK9o3|;QeygZ4PCKzxPOFph5T{%uq)uHv<_)490zm{WuM0gvhKEgHIUEWHxE~dpmy(fv3bAXd@&QLnIYjx_Bx%qJlDFVjj%bP@ zbUEWMhyE^W>x@R6`aSHpD-aG?MJ_|v% zvaTVG;ajs=2<$YL_)<|^DU83dLz#oxcx`a&XcVQ(jM6}6GUdM?>Kcc>`II#e;Vu+1 z>wIEuob}(OtN6ySH9Gr6F~!~XVssWano4&0tg+h*@!5$*Ikh-r)MO>0p*(dcT`;N6 zLay>vSc@2c@Rla1BD)!z-T&okc%4f+z2E-0$NwJ`MKnlHC1mQ7ESGS!o zNM+IYKN-(qlucz~|2+PVDEFlI2kzrl{Un!ICgIBk`|IfQ z){Wn(Un^}4x+ns*r&BS0)~L_Ui02-n?}2y(C!tv`hh(z!>=3+2F6)iSgZHc@fy>pJ zLANEDf#M0}>mO@QI3MHqLjU*%^GWATvIP+M=|e)ihGEK8>H$be}MP!WLpP)rKyTwp1=cH);GEu^30i z=r;E?LhB=r|2M)#ff?s>wf)9{m>N1p`^A(Fo*H+PlU5x<2LvIR%VL=ZGswF0`(dQU zh%Num2(6UnT!liWl1VfE#g)^6oI~mcbxJAY6ACGl6GrH)Jq8*r8{HCm z^Y5_pS7&41wE(nu#nu^{XRLEP!d0G*Z4~zz2|O(~w>ujFquiMcEGf-@(fjDNly(TV z<8X#oM(#1r@7%`Vbk1j&^b*#moDiKd0NGC42XAMaQLERrCQ-fKmN)Xkdiw^x6k_wbE^ z3#H`~leHx$Zo3m#h&EDLWv6R{r0?6i;k6-f7G~3JI85d@OH;Gx5rRGLv6b*6rX>uA zo#lN2#cKSw%YU{eu&~Xfu;^LrMIfRa-|KHBT#~D|{}cm<(uSSIVY3*#)otY?*e;LtBQfq#Jef;@Ty{XUn+Bkc$Q~E=PDkF1ZlqLld9&Z{` znCw8X%l8;wR_>STSndkL!WVYD+L%@4JjrU|7s3;AAai1W1nml_eHt7SUs zzQCZjM2tx`A+kB*D8}kqiD+CtcwNXC+%7;i zPM}2w!Hn5WW%0Q~{)b;MjDIur2+iDZ1c?P(WWHAzPC^K+ttba8y+omZ|T z7__*|b7%MBe>5%(A=`9SMz%swz|Z9u)OH;UMEQz^wqMz0%3le&J;vC7{ewk=La{Wg zQ0ATVHt$gGAvje2*9Qv`=czhsb%ovjX{TguKY%v|4rR}g+PEs{MviS z>$OyO)5XR;4ESZM^SxeW!~K_@-)o9rJ8JZyMXukyNS6X%&R+NN|FPCk2r0Z;*O>1gj{raR@-G#muAeCzfDS=xJr5o24rRx4)>{$CbAaN!=`lHC+bTpa30V|Fk@P5 zi9`6*YC-rk(THwdOk9}k3rQe`sqtiflW_*lW>2}QDmj|M598Vl-^=C?n6qbFcT9(9 zENj9Gf1RCUY1@_$%?%C`lId8>4Y}DX7ynI+XI`dHQa)D{aPf(j!PgyE#-33*o{5kNhY47TA;Qs?|a##(@nHkB3X!7@|EgggZ<7X7- z?T)g#I}`@D>SLV-4J|2=_-NG`hf=x`S<4EoG*3FkU!`I){sWCb05!%J3)QU$gu6$u zg#mfBkOg7u<|koU1mjK7)@PErw7Em8-SLOBp46^A#4|n_^#ZfT8@HU7HU%M(7vKyB zWUl1!4>&*C9WSu+>v9mXixg&KDU5Z)koJ<)iE;a|G2vB==9W`AH+}e;Lq~bwamRco#Pz}jhuCVbnXF@Mpy z=+EsjrK=H#3&vXXo#+A=TvX42IyhNs<8^%>6S#&d7Axz8&^G3U-A8H9^bR{9F0yMVHo z4?=6{r)h=*EWK-OzMDV!MQg|D#G*!x^HJbwo#nBa@Kx-UG21Z9Azb-t*`+z>#(C!U zBU7<2j@l>dpboWKj%BT3vr|C@_>@CMd<(Zn8pb0}7Vm@f6&f>TJ4}SD!@s`=7u`*? z>7lfvT~*j>47zb6Una+YkZP_GdTz`@K|erXd7U2G#42LaA`iW2{67R$hc!B3YG{MQ zP%i!syg_&KBveC$g`(~gv6SCcA^3_Z9|ok6J0AMv*5E4vyo@M(G{zWz3`DU`CVvTi zyDdbW_~Z_IA~+L`DStv9+IaJt?{$LcpB1$7@Hj4X>$Cjz?=eI?V1Z25YX&--L-Azt zqYeIzz_T>$Dt50S3K6SWs>7g6p-a8HS3i-l(YcL_^@|nmZg`xsxiayt8=n>}LLcgQ%{gn)@G7TS%wL(!o4u)A-L7YU_ruH345D9iH9VjzHtR9una=wZIafe(+P4inn=;3ixFp zQ9ziHIi;*w)h)Zo#=43{#i1#My`q%V#9{3)+r2L*e7VNnJHF~|JxP5gCHVkHUbdv^ zTYm?Ko=tos&ROe1r99Tx*Y~qd_&`0i*0)71feNs8bp%Q)H#Smf z?o;}_QIiKwwgK%Vtjn6t&3-D(vaF=}qCZqKjgRr!C-avYJfrM~1Acz8B`XmMr4fB~`_qjh?U*HB1 zhSkZyY{80#b3w4xU9Gn>3I;Hiwf=sPv2fYo{Yj%8LS7w$`pt@LzxD5o8c_6G+i*6& z$-~B}&S}I_6zxZ4=zF>!*_(&@bP;360%-p-1LwWqj@9+dI5z@&5d3bH;4( zO;7BpCke&HVj6?CTRpS6Ak6-%xlaW2)GqY&KzYBw(_yFc?S9F#OHGwpNpTBuHPQ-S zA;M7QdTcbe?zPu{H8bpBlxomb_ktaxR3Z(!MzD2(WeenDPsUUec(Rqc76rb(9Ex}h zoDDe3USJf2$wn+~eJyJ*WUX9xhngzIJ^b~gy$ZDYYg_=68W-t8;k50jWFPfohg4%; zTU?mAq9U`hnA=*d2?euNVdX5fBm1ds?%adp0f$gli$4f82f_UqOJo77<@p0CJXPCT zw(P33?+C*Qqg}I-NvP+@r%@4@v7f5$VQu29hx53@`w}cf2`CjBu#8)aPML0SXXeth zwFGO(#e7)Oq%|X2!AD#7o*1GY^J+_2lSZde*p&uvkyweVQ;|L-*zJ7r)J#yTKSbHf z`_+Pyn_}-*6XUG4s!3Ujf7UTu7&2CIM|{mnwP$B331MVr?OVp0`SYBkbJkO*tI0=x zmdACjru5v_-exDa`#~B8-`Vzb1?y^XBI$ix!1bTQ{Zt>^R)X{8`6tB|K=aCfB3IX-&gj4#c=zm=1e3ji|`tJQI2cAyGeZ7h7N#{T$CFk5UF))-GbSt~5_5sy;yf6CX!bL5c0v{N2 z8P)T;0}tZ$E;m7>(SP5WXAU*GuQtzqo1w`e;)j1R61V~;ovqz$+HXFbyld;&JbPyA zx`JJo-MU{Q^}HtcL9cv|1BVV*z7DNoO?!~VsoUfoAcHSGmPNd>m8EI4w&Gb0Zm8W` ze0)DeDo+r4AA>HwM_sDtl*r+3nrFr%rl;laoXihuj`h8>;e^!Pf8VKZse;G%<=*_ghl zSa`4k&cq9$h7y0ei)tGf+uS}cCs968`X+gB$d@DT4qMarT@k?YgL=)_k%>Rr#M z=3-E#d6N2-qeo4en|jiwI0{GH{4V9%3p)!%`J_|X93Qz=mEgq2BhE3DA{!}C2%?LI z8R|peuaK8m$sLLiby;^igg(mPeT(?>H^9U9wcKGvgY|)?ZW^C_kM@TJf+GU;6c$>< zW2DoXB2{&Z+LfAyo*D5!EENg7m(H>KXHs^>ThiRBTKjeWesqjLxNP%Ex(1^VQkJ}c z)Zdcv_(C~&#zP51@xj_MM+;~fXv<|yL$!d>U zLJ;_suR|6f&Z{wF`PWHld7vke7Ziv;m;16u+w6Y6;xt9S|1&Mv-ARemWK3Xudk%>7 z`g^@28f$oc*e9>O>g~~Us>tlwx9ARixJ5LRcaHSrZm}`*^DT+t8O&AVPglobzN5_e zn`7Q*|I3*Ecz7?r9*UP7^gsI+Drb91iteX9oK04bVX2RgZ_w>l!gA%31p zkE;grsb;k9ivpcOzMWc1l$q0c8x1=Qk&MinjXNlx)mDf4Bz^Zv*G#;&;7wDLML3_A z$YIl(eBK3zmua5WvkVs+l|h%B@W-pS=L?Su^fpk3Lq}3aW=9KzSlY7ta#0_wVgP~oCNerm7YWol0QUA-tgpC^_{lkX01EYxg23&cQp7eyJe3W9 zdlE#Ql*!SnBO?rpMx%X*}&=Yu?pUH6f_ zvf-E31m%}xI#l>_sFJqPD5UJv^jjsp0rhe91#9|S)X19zF%eDge#9m{=w#@=a;4NF zmpomkn6IJMzwH{hBfZEd+2@LZ$cc#Ehr^cXE8%rjg)6j%x$c4KZKklu{l?aSEzQqO zyZ8kylk)m`T9YpDdTxu2Ep)Nvw9~}o^?c-)xs#WZ4*qkIl!`?PJ@emav-hF^x=;(~ z^P!hFP>#?T2pXmuusvkBC3OSx%MU&u@G1F`Th9J94{?5Rht%Y(<``nQ>-dE|NP!(0 zpcn@o2amhq!(z~y;hvtK*9GKM=05@U{ebxhpaQ@_ojDp` zp`f%xbFlsWkhT(RGJCD_(0>TNTDytqX@buZo+{bx`dk$N+LgWGq@Cr-nL%AS31Z*G z+DG43* zf%L=4jBbfG`Sr}R(|4bJPvWq;B{_POm~Y}lWk&>xT!8w3$AZmeoy}KCpPjPdb@sXm zP6L-zzipK?g_sQ+P`t>SCwQ_>YcT8aI<0%H3*31jWOL36vuj~jI*#*yM_~gybVR5E zG5`i}l}Ig1BTU$h78%F5Mez+*UwfpQL791KMNZ%4rI(MeRK z=Ymjwaeq(SirOo}Iu_~c?md;07`klR<4JYE^Z<;)WhKxHEY1K85#Qm9i_j{8pThxe z+UkuB>L&C~z{S76VuFh*70+YBEB0@ADdOr8Oi4uN^N<}se8}TGK?d;U{oq1YuX*;` zQD3x}*M~A(V~MrW{YY?pP`(y}e;sF1IY5x|Q-y#cLjs?ZhClfODKC|c$s z9egDVPK&z^l;IG*99OMaBNq;wc<^=>g;vlTsa>x&=fYYE;h$yNhd2Av7db0P%fu$1 zlZM6;_B%1ehCL-Cr2FLwmE1fAstcgE1E@#e4b(xq6&zj)HV$MLCCEdvh@ONMW$Q`W zj0UIF|M^yoEXrZ>QM4F&gX|sEi~QT7ub*yOafYYf(^}eBT`R*5`gl>FqxgX7W1eX~ z%NqFZm=A4(larGyHH!T%@61omFBukHD8909zr1C|*uh;LkUeng>#>VKs9hAKaS}tfh_0u#5+lI4N+~@0!P$U>Oz!OjAXGoU5u7xqY7?*W2_z3cVqALwD-L)da=|t6}E#)econT{1_23xmw<#Klw{| zh?8fSsa5;{&c|{U!h!x!c}zd8m*#FseqOh^&Z1gYVwCXM(Hk`~Jkls2WZ^aPB#cAU z6({BdWO6usS!12uPz8z*$2zi{2#G#cd0NpCYtA0c`Y+no7-QzZV#hOS z>eLIQN{xj{39Zl*hz&qDZup#M5$VaETH%!Rqb{>7tm>PIT(KN)6|ezXb=H0gtH!A{ zwED6cI&833-*1ZVFot%!9SV5Gaz(oJ1tIPV(lT@AkLOQ|cIIi|8D5&Z*-sXc&zTdU zenYy=_zV0e?HdW~HH<<=8esYFAzF@$JUNNi%j?3Fu)%Rghdt?%E*Tj1gz3ZJTPOJ3 z7^WwH$3Hj%oLzq$d?B(3Fq(Cy$#ISZn|&wSf97_m5*ql4N9ouuz_tlf+KeLub=5p8 z7h~YO^$A4saH3{VFDv|Nc50bxo))mn=)P+Zb$99S3+jueHFp7P=-+UGH*&~dxAbrb zPl(P^Sl>N7&dNE!$}3ZBJ-zSIY#X8^$5f!5HYxd}PO+nhiIa*ch1y5k_vftE{!9>w4r`}Y~-`^~| z&2NY)v&-=3WM)8R6Zc4s6&-Wau6o5pN!9F|p$%DwW~yB{6OO$0Yaf3mMrDhk{DS23 zKuUFKqje;HcnCfWMi39wP^owZBbzA~O&``5jn6LX1BJ99p>#OZw`JX8b$Y?|6ey0U zqR9*5%o=6S3vkiR+*(~C=BEp2IvkqJBql-Mp!E6lzALJh@b==Avr@s2o6yl9VZ#H4 zwJ>UAcA0dLiZo<~2wo3W71W#NtEI~guS!=w1Tmr3uL&aIBN-x$mV};R6sfy3c>(IC zfMrGALrq7PvWO9kb*LD#{NEUtj=o6Vfh+?f4%6wXfd!AUbK4qjHc$xV z3>?6Td}q`HQPh8VT>n749OEvGae}j_>D_LdO=YtgdLN5!x@B}TZOBP}mDeR1T4gpV z?eH(sWj6i@Ox`}9XxfW#^>49-yLtO-EZ^lAsEY{rep2+L+{VJ2eWld+-IcW6=fiPV zH22+HyH2bJ4s-eRn79C0m*`HZuHc%s`;R``1hiJcg)5RKl)Ifh%6CxFsKnieKiJ!M ze$}j?)jS7fJMSP>me*VUv5gi_0(u?&gePfv5P3W2+VviIPmbTecs*LZFXw#g*MFZ)YG>(;?xrN^_>H+DZJiDiXjEVMSJI){eZ%nj}0Su65kp=dM<1q^; z2l#o5Z>je%rehvq4=;_fE}jzCE);95p3S}YpSebeeG-`Q<9obl%8OYZapvsuF`L$| ztH;u#Z&6UVL+b~x3!MBf=IHLiL!`y>q= zkT|gmO^`*y+G7yhe%&Krk6Shgs8V1z126eitr8_U71}(qtP$My&G2*29d~1S`14e> zS&3zM2Z&c;IEuX74+sNY2XQobFJbH8ML{PgEBOe355{ePrm8Yn+n)~v_JGfUKSkcI z;!$tynECjN{BIE0S32np^A4S~I9X8u9r1o=CaYXD2d#Mk-9X~JwAfS#J$#m6*LaN= z;93-zunCtp0y!D?9@ko3NR);mu-5iybA4xH`OIMA5F_Y(Fw8;p;RV>L-g%(hl;YhL z__4dkk$x_}yv|#spZ4!-hgWV_S=={SkY_Y;4AH(POTFUPclu1vYsZACT-non(j2E` zxG`F}7*_d)it24#MPwDIfg&?K#|^2qi#QPzw5a}VijoKUnt*Xn<=Z&c$FalndjW@r zw;*A>_lKnTmD$yIb5{QS!QEpTm#amcciW4%o>j9wHk(tk%hTxG)(IS!7j{}p6&?q& zr-x?}bjX6Nx-Ru#VN|tFYOmLW-K7xC*40EUHuynP&+PBu1G9!lQrB<}|4vt)gtQW&A8IOZaTjMc^3G%st2d^oo6!#yjpmgw zkyp~t-mVXNBH&y!q>7zH9S}yeWn(MG=iTKaClp0wgB?(Y>xB^q zpC4zymk5I4k7a87w67@?6o}}s%AmPn{S0Kr4_Dx?RY)KCo<#+N{FTBSh%7)tYzHeM zV2jy=&9vSO-T+5My{>?P3y@N|WCNs$^M3-LMVo7?C-ATaw^!3ktJD51LiYodE9hml zCh>v`1p5rkcsLS~=SVg-pO|1HM7TGua zj|#Y=ZwqcZf}5H4Y;e5nTi&sgj4_AVUI!UiWEb_%kU7uns}@V?|E^*nPu=?ij08k_ zt^iKnkohPZ>S3-h!w+HCiUS3r?4KCNRgPU3FouYld^Ve`3hr!A5^YLict0Lpe#6%h z<;xvxnuriRjS!FCRiNdegKl?UaTU0|G_gBBnQz{nfoIAhZOrR2r($A5?qnGF#WEt1 z99Vx@(qpkf5h!1&5!pP+KUKFv(-URMbz2Qe{vN0TPD!c%@r$lG<=&ndwzTxYz~bq! ze4gMBQEjhykEzQ91%tU9u~qArW|?nR#VKrMpPb>^#V^A`|1e5`hs4#eljCWOL-}h? zvvz$c+n8f*i<}IXxm3>8`CXGBZ`ZW$mqGisV;Xok!EpS%x^@_1@jyC%4=9yD^&S;^ zLqH>QYG$q8JZWVs^wSm6$Gjk-*f`E7pMQ|l>^e2Zj}RO529ffyGa-$5s1GV$x9Fqdu!w(BE@NzvnwJ1M!zml zzBQ-ev~`HoH1+L%;F0@L_cBWhQ)~aGVeHw2W9?_krA_Ev%!w z_bl|uG1)oOhkn5uaF?xt@HDS4!N45a7_Gkt9B>f$r>_c^Vy^uox}xpP#4%^fX*^Kd zcwv~-ZEe6-rE6I6>S@aDUIFUHqvKrn5Zc=KH!ZPC{$vi5spBpS#D+`=p;n|CPADwf4 zqzlnlp*P4LW9Zt2zOEJHm)t8$mTI zNF;N(dLJG5DR7nKux2pazsv@@N7zyiv}0wNJ}k%4&HLq`0+j@#4T_T6aM_X5_jQ$i z{dmg7Mc7RKxFpx=_Ax>t3+qoT%atAkNGHt$K$`SJ^|8I9gB{^FZtT^`< zH^LYvR@VfQokf#oGDBDc=5nUUG*f z`#{+Ijdo~hUFT8ixCn8yh43QTw5RbaPEcue_uJFeGXz!!{g~x4sPb0;-z2F1u1?~m}kJjuVBHV ztG!Ps@IhTE_RzqQ^-GK8dNcRasumEpXa}dfLNS$f4WFxqOx{4#RbSY6K)~LY*8_>N z@JPFzmePtcmzj9MX4_ooA=i)VCIjVlDS_?n>ln>h817Oi`yAtxwYl-4`qARfw7Q1k z+^nSn{I>hoEh5U#*AFq(dX$gD0e9WyRqHG90viXf&jFf6F`>!*Q?a7@-kotuxMJ`V zRJ{B%P%^LI1wo<2j>$L&fmk?094UF6@lUnx_ptV5cJeE00Po2kc6xr3pM;e$jC3WkXMnmfH;Y@svTxxG z{pH+w>69(4%%6=s;DJ7yPYW$s4$JF$iwk4J4ktxv`X{`JIzFh2SkY}FPc!kz9ug`j z0Nj^AXo_uglPjE^3@{YnX2~R;&*rZR`!&}*H^}zw_wP?%;W6;!k`RmXCSlrVog*k8 zT(h8ZvY>uCzHLABe}aRBQR)<9yq?PBnGTZ6>0dWf*p?#%Pz3qRUJH1#jkVsAlYo`> zO^W}^0{DA6P5S3%)oBt+Gv99&DY>p`Yy5uAc#QIo;>lK??NrfjAD0(j8tLBMRoB_9 zuiz#LRGOCWVmsqr;an!Bl%H{F@}8o!J=nJ=bkqEo`G3b+pE?7%LOkj+pp>9;H}=~fd;2c)u8mIy<{xt5tF9xM1S2} zhLI~uaX$$!Sp=Et9(1V2rpp-pHuP*~IfA7N+dUzqs?8|MhHm=(F`LnptEFoGR}%Xv z5mtv9KZ2fk=BI&AnovXNl?+2EQPP}&x#)or3&F`fAUKQ=o4-=M-ZmKLBt7AEPs+EC zOd?gzR`YlfOWG(>-A`_sEJkVm^-JCn*&_Osr}L=&|lTU+64AJ$_9Dd-+kucTE#JV3hUe|#A_=(4Y=Wa zevlxW$$=C%2abf3`Hf5oJ~z3ar)<`xbsc@qbyKz&O61tM3wp0-^}A zYRY}PhGa@{(%R~1&!ck!P9APX^+&!Yc0R8sP3lcZ5w!CgrlV`vOK;?z)G~f0z`q!q zVM=Q7oX+Cu@lH(5y4o;n8aW`rx0M7=|ndWqjYd*bLE#{L$EeQ|; zHU)6T_8kI%_t#feLT8Q@KS9=fXD)5R}T_1Dw}>4eTPhgp=U zvv=9&gQn`&M9R(z5hn3S+@dD6|JYx3p){Z6wyZ4mz>UYCj&|1#)P7LLP;nQ#CbnYq{;N~XH4{S~5qB7HDqD$iK5rq6eK1UcRnA#wz$5` zYNN~bL7nw>9BY`}8o%9t5>5)~VtwMuvXH|Gh_@lwN3QFM0 zI+Sh63QHFylPcwh8we1>vwYC*XE$<83x3?SyccY<5cCIT()ia4t_81ggUa*Lg7cAg z#iYEL^F6qyjMtWoy5gou1<%bnKYDvc7?80$P4f8Wo8@`*t_pg>XweW8^>GLgp%NqV zCrY(Gb!}1;Dgo%5$bCd8Y}{97=f&EL+LbrDlw1yIBJN735OX`X?4GI~I@pr8`Go~7 zC7Et?V9K?GmVv=xg!J&I#Q4o0080O#Cj)%KJLx9pxgP5_o$JBBaLLO*pmK>|!;-_y zlH9}6E6;n?IoDZ|V4P=?;z*+|u4ZK*RT>RpX0~2>>bTxUGjDV+pW`Y{o_`!gH3O* z>P4Cgx)MoY#w|}&a9RN23dWD?siusf0G8%i#`+O7q_d~K)SimGqN>oq3Y}fNQsogv zKc|x*>Yuqw&Dt}f0O0COTyNMuhR*NJ0+z&l+dDh`#z4Pi#@CZ-pjrb>HiXdD1pnM@ z#oa$Uq}c2{ZfqZ!Y$1{(rloI=1b@wKX_PwwI6~(^&Bll*BjID7q(tF(30W4&C--(b zpt(KU1Qg4m`U`BbRMSU$lF6)ItPV6pUhTo$CZ*Zrs8nCXi`>Kj5@_P?AL@O*1Mcr0 zKeO!Wza4^}-;RNeo6iUFmBrq&CVsiCw<^19dj)a4L;knjBJ)EU-=~Q$pSt3+Y{0^f z)4U#D^f#@pQ~6YZ3w=K&KE zXb9}^T+rU(>bt9iu*i45^bZzxfMz#xiC1(rj>Fb^x9v=oC1!3xvz65rQmL1MiF7`CcCUI#glRjpk*0+SaA zZAC>D9I=EUMZAsdxZR#zPaoVLd@w7V5Wlsm(ZRSuiK&p$&N#$Ph=bx zWB^}hEVx-41F%i(2yQ* zTb1RoM2+4e2iyYDu^EKRD|Mm+D-j}h*DMvQb?qO!j4Bc_x=zCQAz1MiF{iYXLBG=K z%7%>TOe>dob)m!{)Q_%lLr>z4(IzC7sTVt?J(YnT)nh^5XH^ll!(l8D!fpM>nc5$I zOs@#VuRDmOvW^Us2CPZ;Nhl#p;mQ7`a9S@0NOp&8m?<&RB``ZyqEb-{uQOlfW~o@H zef?%5jX-C?Xd9WFx2V2n%bQi-23#hN{YrVBkk03yN)2C~PgTt)o8I}$4tAab)mvLQ zch&?N>FybxLz2h6nrRs3htslnPd@u5C|vmVlgD%}#2UC9shvcM&a(`l^O?)s_?$xm zc!L;0_yZ3p%ZRdf`_UG|#{4mt5=3D$U`}ToU0V;%_(tC*9z3hrws0B$09~OcoB`xQ zyl@XrikIXGK4W=;JEg(S3L>wrnL<9zb2?p+J{{Z1UtFM-G%0P=YfN63KMK&^D(LvH1fct$qg?<&G}(6!F@Yo=*VH)Lbb6PUdHGjN44Z-|Y-PovXZ(`=3s%dn!? zpY0~DL7r^#-ecGM3bx(jz_hp4iy$8opwRUtWy_J6y34x-ORJ?IDBB|U`E$Q!;LhU> zuKlMI!5pKHc+t2_+;LT*vfWP~Xvbo0%KZ4UATBYGPM;o;_i#H!@_)aNix9oNB+q9NUCN_Dk@a${LvI{m$2Iqv>b== zT6{b5V_ zPBryVThGQ}sO$15h9Ge1ylRsGNW4+oR&%K{-;Mnt1!sn(e#yzgg`zlAeqPwMDER)1 zR<|%HVfU=08@SPRQ}Nyq0J zB9n=gVv}Gwf~)?aN^7VY6Z(XC2h#%~gvZkNhqF@(W(r0*gQkngU)zK9p7xdEduj;- zykzdAidx{ygyj1K%n0kmG$-c{mc< z3WU*pjukz5BXcSzl0DAYJ49iBYZ^WIFrgN~8*=PXX{;r+l_cC#vEJ@`Uafo+Fv zH!&Q3*>Atv=vx|vXF>8}oz6s5Fu^ckzQLwT*`-p0vI6wS5a=llx;uMV6XriMRJh zvVfM;X@iG8ytS6gf9>6m7P;0aW`=_E7)*zo7E z0?iI_qWVvE8Mi99TE3zL#16Z(`rXAkdB89EqnNR{BRI?!$ENDAw&LWfqM9L0Aue)R z^R5)f2z|+}omjK&RoLX!=tm(oy3Dos36@tF!^Fl-cWMiZa}EW;1=mi36kFK~pOEHF zFpY3Ab{~fG6(#s(v7J@=nA&;5_fxgs+) zHPmcS?wEHNc=%-h%+EWf>@R-CZ!V1h%Snc`Bm3eoNHQ~^+K1{&L0!XmlOi9U+tmc5 zm_e`Kcjmsyxi?a&XJcJ4*jBygJBHR;rSWc*D;pZDk&{U4YK*M7TYNIo+h%vF?_Vfp zw^CkRQM2d9pAQq8lBCAWL5k+wz^{#Z3PhAA((iU_qDB4IlcBTXa^{Nm5bjSJ`#q$IIG^%nu2MX;L0MBkD|c#Dn@D2;Gd zLv^q5NH={jM8q%BU+E(NU9b-m@6LYb*_j0B0yy22STW%Xzp`fCE5TjC0)+FQAhArR z>aKrag*Fb$GzSK8Phjo0eUS4n8Dc752y8V!CM{SUBQms5IY;xkP)RB zvc9b1$Ckd)W^v@W z*W(@0o9vxNmXPM!Vm#{#7t*GG6}c-@tc!T(dUHrI_I!h80-1_@}YXt_U){4^Q^PDA<5E`pPGs;SoNw>{Kp_h zRf4CBOwVTqL*m%{uey72g|G+YP!)sp)wZQ|mLC;VMy$;uTa}N?#l`!2CEz_qf5R`e zxt%f#6E+yNIuoXhr+Eq*)sY6-v|2Dkz(fm_b}&hn%(9Q8S}^~N)-D2A6)9Q9CRi(* zXRFL!*^RYm`?++==L$*ThL9<@%dMF>zu5LTP{ZZN%btfO?H8)^AHnj>)Tq4+z3znHY4-#jK30d0sHIW-{{iqo z55K&}RL6+gFq{e4sC*%3!i{#S#MJkf@Pi-xz%IDp0{i4AKk2%RDB!2XQu)R=zF|Bo z4FKnbJ_(! zMGyIuQ%-U1AU}aJXqPXOYe1`SA7CC)ee(V4SHH5`Z@=C3iF`h|d;Mm#{@_m?r{C{640>@DX75`}*LNiVRkUj1de?H@~x`x|ch zz3Uf!z-sktUiG@Cb?3ol4>%W2SgA%gc&g!pA3M)+il98dOy9|L)aPCfmpG zyU#gC%V6=J+a6X)>?miYUxYga_aU)utTgOw>ub>BsgF=;3i?R(nY$;x0mM{DJPM1G z1nwyh5v`ndm5L>)T=#_KSCfVADs>)H@deCtO13N`y(bbk&1|-Oby#{%d220H*@nSR zYn|0$mD0QFT(Q>aXt1+}gYCX9#5!a_w7$|RXTvZ^8;A z5*bt3)y>s5xxaY`7lrF5vqM@ZxIv=CjQ8%kdqLouBzH(5|)mmiU3RuOIb;TT3Y=Gkfy zN&21hqZt`>(e9zREz!CvOD1AgKdseb1G1zlT{7OAaHYiO#5HM})^0-*#i?qL&&`fD z?TB{SD_(epz3wHiaqrG^`Dg7P|6tvB|8t8b%sl#-NrU2CM!&hRQkRVt3d)wf62T~rQA4dW8da^esZlRcKBYA>tu>&o zMgUcAnq;ZER=}~kUh}*RRt;&&6;73&nt-l=^k8jD+$M36#63-q!F2&ZqBkt?nH6_m z;5!sbThhd0&1vnBjxN)#Xiw?&wb{R&c9!*Sl7(Dx1QW@e4F%VzT~X_)(;C^@tn$nn z>pMhTH)wSY%RtLo8|>X^16}KEq+hy%i6O0b0SB%RniaBjn=-ax06j>E;iLn^!HiZk ztyx&VM*o-wsEOkBs>2y^slaC{Un?sUZ!SB&3J_Tu7; zFLpkEIm0Q1Ls}XDH~=Zc5TA*wT$*ohzD8u2@dd15!B{ zFAth|hq;Gy{KzAZI4&?8WPVq`st1~s&oyOoja0zu)vFzVq+Wm|2b^T?0i^s^)WbCg z@c?PADTjJ#8+CEecH;d(+c4yI{`u#-cZ&@YxYXqH2W{b+b7`Qh^n-)3pdCoc;A)II zc~1E}Bl!HZ4e@gy&lmurxhD^G?akqmPk-6o`mVE_b!j-aKH$gr>MwuO-&#r%KDXa_ zo85BjU+h=E`-Rgj4v9AJ>e(c5>dAv+jDI0M4IO>W72mcKk2}fPpct=M7yK68yjM~;->qo4!hLA$QUh#pSPgt81Z0RL zBf#c9E`1z&>2)tYB^ zNY^SOFg4Bk1>$nrSszQu$75~SYCE*E)*&#vOvFK0JJCXBiGL<+q*sPY8mqKBEn-90 zvHq0JpMRjm65{sB_hM)?Wbv+kt6f-cv*ygQ75A>N%3y`{$A?8|R$F3d)Ea9VZ9o%b zRyt+@5k?ReLmG#ss6ews*s`Jt8dPUe;J9L;biSnjH6+j+3`Q*0-Diz++AJe|w174Z z(jtyB60fV$B&KZ62+ENBT1GW#WCfC=8tTJ>?Q(>OI*%iMO`)ZeW(r>$N9H=m_Wi3g45?60@{&4xwv z&zL*g)^~MU+w5r;sxH__&yc`s&SoAs%hJ7R>sZ)qE7q>IcDWQ9Lwrw!PUOj@R;dXr ztVykFc*qua%(79{5lBUB`R&UrEU~v*iOVI1iw@z_vWOa*Cb2zDn019}n}5K3>lx@3 z*F`#R+AzrLHx5dyu0op^2~8NaO$`>`G-LsR`!TsP2~`BN03{tUI<2<-IS0mLTC|?} z@ekd*?*B9|SZIq+f4RTi>GRNkZhGMP<3<5PM1;6F!d4m+pm9Mx8LwH z0g1Rau)^YCi6Eaq6imIAU0VX~yTq%of`~IL4!`8J=h#6Z(MCXj~BU7 z%YX7CJM4oOOuD>%DPvCHw5qB`8=JC}EPx6yhC>>cE^%5LieYW|)oRU1i_1GQ64M$6 zktm{9b7hsd#c9oBTH=&nT+_O0iDL^wYE5uJRdXG#n+lpE(Jhv+p_s1PFx7@(Q~(ug z)Om4Y*$7E%Ey?9%33g27X)O+G-W-!ISu#>-F&T1!8& zliBnp6gRoI{GH)EY27Y~B#*!MgmNh%T#WN6N-}OGs^sZDcR3=7fmgiZ6>j4j;00JA zpaTp*hRcTt^flL9<3QU~2jGP-e4%sIL^elH|eiy*^+ur6 zR1&vn1EQrdp#@b2IYd}SwNOJGq{S6uXE`Fy*ciJ?y5>V7y;!#TXcVuHBP8M?BXFEm zo?u9XgFsD zp_hD2246JEY8WI^7qH2Po>lJ99ws^L6_Xq&TH7@aPYZWMuts50lG{f3z1h879 zA0uL|tFy~`G}+FZJZ+)$Zz9EZhT`1SB_KfTM^X4a=^>a@VApbr?km1dfz?PdB;CAwhh*H$f0GB zJn7Ngxr}R{TxLK1+J$a2Y}1(DM~Lwvt+I0kJr}8cJH>75e)0k9{K}Jd_;D|?H=h5Q zNwua=#Fhue6tMB{x7y5?zRa2rerDHWld5np$|Ni*LsTip70rr(grf-!8PHrdq=!i4 zWHn!5rMkLWE}aA~W23zSIgJ9t+N_mWb$|b`^vtvlXx+$#M;w<0OVt(H0Mzs+qh1-U z4_XK0<|t5*2(}DpCAImR5h%_Ia7&Wg5w8Knm<6f}+VIS#MPMhz4bof`i;JVBwIQT+ zA|=3{Yson=%vsYH+1%NSwQ;)MZu#3yBF;Byt*jD(K4`Ud4c02I^_)42?5SMz zvONLqqV$M+4hVGZ_qwn}n$3Cl)zghP-sp5j-}k=vZP7&l#UK9g2lou)Tv+@4^rt`V zfXNmKck%|qG=R5q_}2&kTs@!zZ~|}vanQN|NdPNxY+`VaXfb05C$$L814N1E_s6&vBG-s%b9hip`G{U~CeDL( z%H&(3P1M1r3;m)$NCi+H#=S~|2gJmcf?}>|kN-Bg_TMvo@ZT_X(GKcjoQNka2mRtq ze|$Nmb?=SW_%A;9?b~;MhTr1PU3l>>?;h{uzn}3RhCgNhTWqTm{2lKrW4^6EZY}=8 z3oo>jPCCiiY2X{0ObAD%hs5B;^#Qw`Cng5%e89Ox!YQB+03Zty$q!K_3!Mq=ibpj; zAWE&T+;NFn3v&Qr?SQ-GRC(YO@arL;w{+!}Es z!?M09NNSNLiMmvd+46x^Hh0l{+pu<{)AK5b=xABkAzx}8&f$86IKy{e{|`Ii)hB5J z2wP)In{E2ju=R)H)-b)p2A0d937pG()Q01uHg~}+S%!Sfh6e=7#bq54w-gcTuCxp> zs2yw+Bs%+LaZ{o*fPHa>*%(Okp-XHnFQX!xHgs97+-cx$BOwtyd|1XL{t~LuMnPI! z)mni<5lMkYaR3C0BkE{c+-RX|Sqjwz3)eLv(bT=HMTX{s^zP#N}Po{0pT%C zN&~T8%ZU&|U#`EmTdq5_u@DgVT3~!y$4qPNXtpjuyuf^ZW878@uCeKh=i2fOtL=yb zmq_$#we=46N@q;EIGT9d=T5hQ!C}iL(;A;ziK3og|9zT;@*|E=J&tV zJDjzJFM6@YyK&0Rs@GyVTkqOz?evc5I~%^Q43eijs5?-=ie>lOw?6Ye>((4J!pFUt z7~g@cdt8zc`XF7+`(`}r7If)*P8{h>Ch4qIrzO=jDx$Qxu1y4Tg-vUnW&io=v+UQux!yW8hsQ-Yk7$lf zNXL26;sfpOzu&HVwO5+}5p8bQNJ5D|$J94I54~b)(PnP3TMQN90&P_S1!xB70gj2G zppDWZL00_GNW%I+VB;L3+JA$t90#W-ma^~O15n`jZ7V!37g|?dd8Ofg1mCpYQuVgd zZ!aEjcH)UAx+GXDMu!>DUCO!RH$ZKRI`@|VfaE;{+yWqZj{$ca#B%_80k*^e@E9|U z`~dU-uK-g3DdR_4(lL(Q0~RTp=g`$WBR_}#W~hU50It4H#-F^t3<&$@KmYmeXTeJK z`t|FbC_DfOa89|D?>WfiUA1bJ6VpaSAD~LxNkg8Fjt=Mk3(mIxemO&@`*xMS1GK}Z z^YPp>20SAl*Yt~f0)6qwYRBFwo&N*V#aGa|JoD14f=R;xjV;N1$8e2xY?R=E%)G--AL1(xB)LM%Oi($Ex2r6C2Jn1=z5E<^4zHmmI;nBDV zm$0=j>9D6BSuP)vE#h!?SY1%yczwT&plD)hX|T>Ut0X$SNB}o)8#iqbp`slXi87}I zVzu{YO${x!QC#SnsyZ1c$qHOISbvW|rLMO@?I{)L*PrzQd;FTFF2F^oz~xqn5kr7H8*BAx4?TwW}kyGVde z6S6=&oZFQGv>}1;#Ro37LNx328UR#x$|;JGwr@us((nL`}NjP*CI|x&iV$#6_*I%a~&$` z>TTh_y~?J~n`hku5HpWGZkIQyUf&i+86@xWq1~duxFjF`@5>*zP10$M!P44C!a7+S z4y=B{es=X|?G5k$)K2$m?#Um$ zJX+Tz)uXA~Vr-CV-eLojbyCb$oim2QOd>9S!lRuwtxXa)+>F#u>YA%$${}e( zgZ#8AAufLT^5qWv0&r<3z!wlryQuH}`|o!gTp1Vz2z`C#7HDEh`Naxe~vrkUB z+lu2EfwRB%`O4?XkH9Knn0I~P97zq0+FkeD>5P9(EREk}NK!)m1ERC{(|{N?I0>m7 zthn!N5RNM9Csas(7S3c?JJx)t`kckV(=!>Z3JMItHLa6uJ0EL>M43|pb{G??WycJ| zAzH921;}y&O6*odG!f{Fa1()6P^*qqS-*5@`A9@sScQObMmyc?SQmtv;4U`+wgtS= z)2bABjfxu@khtx-HC?t*R7j*rpjf)J8`f;l4t$GrVuxkWDk-a<2MQSGY~xA+Znd+i zs@?`V*IUD)b{k&SCtX=NIG0HF!3!4JkoLmk;=G1Z((@I^c4PN?YhO6m8pQcbC~u)s zZNGb?4KLQtpZb(Q$4rJr9QRy=^^kVqa_U=9?XAaX*E;EPLC}O{f?xW-L;V8+jM5_( zXRt!Yz^Du;h>%K&h{%RTOv&(6RvQQ*CJc$?NejdpC6YWQeYIhUkhj&%lq(EzbM+fw z%qpNHOAbHS9=-oji^@<-N@XO1fYE~0iz6+#Y>|;M8}1Wj> zbc>@=X`3X9T_>NI58m;BHON3qh5SyYYsYNUx-Oe0E2iOC(7K*nW!VJ@x&M$sYIPe{ohurF(tw?89tz_evXhqEq9Tm#Fu2 zapzMuQ^X>>mpM(!J=#z}9KNlg**3@sSwx^aCL=H*aV*g1i)umI6GE;}9CR%*vvvb0 zp4h)KNS-(%yYBuEKe)pB#97;yK(Qa4bs4z`y7RAg^b5|gMTZ~jqj%_x&U~X5jv)*O zZVsDn|GOrMW4zdVo0%}`o%wf0f8{%QWz+*d$sSw<#;tX0hQnqQ+@8s8tyuC>>`_SN!{`*letX)@0s<F0%k9L#mnrszy6EbstUEp6yjM+YTZ^Xn~7a$ z02-tX0=E$$@Sru$36g0&4NKw)RjFbM*KDQrK`5GzPyz{DDBe5Tg0+wC=v*8CU+3ne z``8JtE#iGMX3QAx*}Qr4cF5l*>*iDfAnWFvZ+1G(fV1s|zy9^F4v=jx&A##s5DI|g zedL;V5pW5B<9kTw9=9)0xDPAvMi+ir7QWa^+jyla4HjBfqyZ+|P&@+zP5s2}>B zvIx}4vAy7*Z!N8Vwzd2`qwVyq_?;K~>94nzryTdpBQSINOzZ9IvEh+nmw3gh<Dc!IFNoPi}Y?X(t-~trXJE)Mk}RH zJM4C=F@93t*lcl$;`YWjS%vh8v38jW%cx1c0M_b_5^-s^u~97ORoJ8VEpuRPNJd5~ zC6a9dbvc2|VOc`XRHK;QA@Sd=t$l2bEj;J|8xoPSX;Zh=N{>0banLe{%&=(ZfDB^^ zv}!XTHZf!yGrSyB{dN>Xj7n2 z5P>8;+-RM2Y9*^C1EU&1M8_pgu8oJRh>xIl?&Da&7Kc5m1aX1iSqtXaz4twA5z$b% z+X#y@fvDX`jda6g9W^cDDx8&=@_=+=#P`mtFAeRj+91Furu2%dq+|WhCp4k9*l^c? ztVS1X49>g6^HSObXqnb1&aTAGWqf5w;zj+ddab2iTwW~#A&L5hv{8cL8Mng;n3i1M z$Z*nHXSP{f<5DefK7HXdTN_?)oxRdmmRMmz@fORAXp{(&Kzm;huD!;<5mo_bDZ6{^ zCaXcTP*%CMe`=A2N^5Ct&^T;=qINIP?*0RTII2E82cZm-cio_m4EEXmxBk(3LPgHj zt}A)ZD2XqnOPu4s{q~1D4UFkQ$6<$W0gT}q@9)4ES~Z|e&yY5Gu@W8BA_pJ@1Q)a^ z85V~ZXpGJvzKu}{YHF^N!K+?tZmE|g>!R^3q>RF1yT-@lenoR%Z_g&nNl&=03XZYn zaV<;eCub1XX7^KT0KnSeUx{F&Tg)0ONoob9ksyssJ1Rne8DK26lbm#+Q))+2-UFf) zRZ?#nve=01Vu)MdT)&LS6--^kW=fT4^|Hq;DK2tMQavN_Apz%ho2j*B^~#NshSIvB zP3nT$%UZ@1jJUQBtjR@DXwd>%i{VNk+g&7<)j9}l25u{XXP(9=rdL|U(KmOdZ~P-} z9}N#Y@PN&qKi@w5;SU@7z8>ssue*%LvSrI`vZ8>cWn=03xZ{qq?|%2Y&af9;$*qO= zyyrbm3>Q7;eLe6S07MyQ4uB<_2n0uy>C89I0Vfu^4nT^VC%`A7*jQ+0EICMnjxFB? z=_ntXnzG3YN0)1ydT9@|J#D9KUl-|UFKzIFYtl@N_s==|esJx-ajwZnAddW{^|%M+ zKa(B-e#8HK^_lj|-(GL;I_Evs+T5!3X-j4jzwz>wPuY)u{(qCMXJ5#OuXw%#j2bRj zuwbXU6b6L(-7!EcA0jVvT%fP`LaRKj(R>8eGL*zeR;@*ud(!H{ae)0t5Yd)697go3 z`0yzkqg?1xF=>#835hyFfJ!I(_*iTmRC8g z&2n3W#b$Aas}fNe!0NLVD+aBdiAuX+4Kg&+B+!+SsIy7zIi`G7()+c5uEk{z$!OO4 zP62EI^mQxNSkvrTmK#i3=emcjdhrZ#00wO4jCQyCUs&H~)oPO{Bqu^nYg5T0fX8yT z1N8!>Rdi4B?2rlY6zvU2*Z?!x!z5^0{WWFC2m;Rtdgo5rHh=k zJ{eGJpFPd;wfY&P>r$iWGapAsJ89h8fnV(J78ZWP3uK zWY({kq+$YMF%QdTLsSw$$-%g!doa``Kg8O=>>0>QGDn2&n0yQCC#tWPJE&SoU{zuT zS4uo$7w5EIL?lHuCcqgGH(FcDS|7DZta+TNsGyC~pzLsDCA~DphGkgqwOGu-u;Owf zTI(dula`C7b4{ENLeTh_^HZjAtVBUPY=kBMhMpmp_qEdsBjf$@I0|K+!h*h~E zfo<(*3XF@8)tb!Q?_4%XGAS3)nk$j(g4Teb+F7WS>Qsl;czHG`$c@C9`j$^<9oNE> zMP){)I$tRWuT5~ow8JV7J!{rR94Kofbx((ig!4EN4d~y81H>wpQZan=o8SD#mM&dt zZ-4vSx7uUI1QA)!+1qNO9gGA(p`#3^@t9+dabO$HXK8>&n?8NIbImi#@nsPQ{R~*=puHRbUE0GL&M;sY9ddM!X&3Rd5nX63^#1O5zjOJ0dC=dK zM?L=Aqdf%rP9Wa5-KU$F)+H?YQ~JC(e-t+yO5;nCmoKL<0&hO^YtkB`8vGaovfVM16&si=o^5E7*yAd>?jq1|)<80Ch6ZbwLj>tS`tSTs|QBJu0BvVh%VH6j-3D`i=9^`K=BYPG?8R$ExUi=&$CQkoC}f(h-a zW2_999vS72$VVlvnQ*zmq*)^!t}I5y@S!ZdfI=f_GY>D)pA6#R#Qo0tYBEmx42N6G#N!P5Jv6g(b z#Fyi0zw`oU)>(`EJ&%kHTSIG$``%dJ2$M8qPHvuN(4UEr;P(~*VE{n zsZaeQYJ)6%SGP<K+YMWh@LMWYgzOQ=d}ur6(>&lMHN$>lTpcbW#&$%=E&fT#z31EUE#E@iR>BMhb{iMw~;IN&lDy)sTM(N)S ztzA2wZ%7tU_h(?7luIPmw6NBtyufZjV$2b3u(E#T+$OR%?({xk4Iq4${0&7)O8Vq)2u^tVw)P^hXwxo$Lvz`xWE)R<<6pcxeC?Rp| z2HE70WKj(qEa?yr$D~3ePLZ4m4+z)}$?YC?A{5a9_fhV_hYkxKsl5LV8z%KfxK*NF#a6S%^U;|Hn}DpfjP(1&(ud6 z?x~Y|%Ay{)y?|~uH@y2Mfbj9hAMbon(+={G9$;|nvBx@*YDZJ69-dKdDGkmw{p0Kp z-^Nn#?i%_Ge<+7QeqRUo^pD`5yPNW#NsYk5`HSqD%fJ1!ym2wQ@U48x38!fN2-{U9 zewCk=ct44_;DQUB3d_eo{&7RRdPjppA%H!MLIG$KGU|b@Q$&7#^IF(g7kSCCTCGd% z8cVE90Io(p5fOdn!*;BEeGt6-?gIxbBI>0NkPiWio&aLRqY<6v8nNv<`PhoZC9aLQ zwGP0y|8Ocq5nyO5t+ZIO#_&f&%M2N+k~=368x{DE(r$D_z|(!i7$k{_OD)l2xrr3F zb-D#5&YKXIv>_xhl$HI~6mGNrOuxl@(l&jLM2zo#NE}|dOzIu6WJ{x5BMn>S^kxgL zmVe{sMkl705Mku>fwcoJ6W*qdtGtK}_4Ny%rk#as{T20srEN_*B4{v>g3scNCbZbP zZUIWU@z7+J5LgZhw8urDMAfg__9jbr%9ovpjZmb>NlhiCqcXSMDjplr&b~GsG^Pdl znUtMsi>M6+(fM)dpY@N*-IK)KqXNj8by?fEM!Lkcb#faeU1u4yN(Q7K9ZO4;Tq1Ts ziG0^fBpPnCxItt3-9j>iH75P+8X4z0^7SvUCvLvSde-Cwti?5+E`1e^Nkn>TQ5pRj zO=%LArBw{f6a?V&@{w8B(Pl}#GgVM#{6^w(j7m=}5U91Dw>A3I7p$w9D?AICW z>k^^9V`)jnLjL~GUSltR!@I{bUGd3xI->pRPrb)J`SoAf7ry^z`}(KeYd2m0^o40W z)3bcRc4Q2Rnn{HE88f#4#<6v4Rm$@xQPRGVB_eLI-UO{(B3K1JAq=@pih~>y_|?lG zsU*#Vg1s4mPVRBxgjjW@Bx$l~aeh@sZx?DjlAh5<-{7cJQO4x@MVpL+vMAm!`wN5p zX=`nf)!TSM>mr-a5_hZ=;VsFWXtmZ5&HKEMIjx_XUcrxQm565@aEis3RlL?$9mrcN zsS+eLKMNRZUCIl@rv=pUnh*@sX|0Xg;IP(l8G*zoVoKsY33*7!sVqkXs6^Kn85hIm z2a7%4DKrTBFT!4N4UH8N8;{DOwbrz>tn1I6CNBpAGHhAxXn=^=;dmX$jEd7GZv&|j z#z`Y33DunHf_4bY8aEPBL5*3CdRz@`JQN9M7qO53`P8RAwbdRTc;JCfB=?a=9vKH% zTTQqVk$|b6{p@GiPJ(%?X7*Gmb z$~bxvYR+(P35+F@exyT$dgI28jwXh|W1k zqdv-~EtCxi=YVSrNahR|o8Omo!~xQ21JVF%)~s=~HD!CSOS`Dk2ioK5Z{jJFe)-_& zY~|ltD9z`uIs0q!mc!1Ez^E8dF}-^UfnI)}`EwVzLfTr}w_4cM*IsVtf9yPa#c8j! zSH0rZcKp#NxO`hpydOmJO?==3AF!*hzS=Ii1 z16qzS5YWqrFsPB3u$uq`Sov^^u9*ai^n>x@``s_Zq5X%=2S(Y5R*#{_q<`!vqr>cU zmI1}=&ZA$Ybdf+UKpKwasG7J7ok#FjRoJL1Nv5vdor|#n39X7jSKa z=Kuge07*naRPY*=I9N>W%E*0Cg7$PLWdN$(882D0a*efUklJOH6F-ylWGJdfhPkpa z5f-$erON6%1ZHL6icRn_wW%PT+MobETFz+1%7KZ4i3L-j7JBD1Xyd`WEGaR>a29uG6G<2vu9XFmQMlDH3H{J zjX`i_w>Az0kR=-F#sfoTOq}{zA&S~sT8j&(+$B|tAdBM*S>szF!&_sjHzCe)jlgz= z^r@55)2{6h*p)Ttk*o|qh}OaIS(VxzlRk4n`9`|>1h$!YrK_wq)mO`Zu_ncsxF@)S zGP$@ah|5sZP^}FTiQ$Md?sg!ZKTnLnXh&0=>BgYU!fW@xMzYi0kIMj;#zipLGSa)F zpKwwJR<{t+J*6T|0e&)$6y6$`WjU?~tr|YqwRu{c|3WSlCNK86KburM3D1Y+1HSR_mtMkv~xu$ zDI%>U*--&#ZLVhIH$EVDcUl4kv_aY&i!<7LKc1WB|;PY?8i9AwMPnT?Dw9s5oBY%+6?Uu#n=b{r=lbun#?8zT|+8f!_4AE8^9}aSsS}w5ZbYUBIn{Rz-x`b8Sfr z0A+rlJVcs(UPP-oNbB{E`E83db&+e!HS-MR`#>7X;!ORec>v0uBkV=1;TQwl=gpgE z4?g&y1NVqn6A-Tk>{Ff(Tysz_f%%SudVCz<8$)9NfBFZ|reCy^yu~?^xv2PHicHTQbu+6;qSN_@ee`VkK;kWG4uYN%~ z#y_;%ZoE^z?dNSX&;Ib>{qKLj9d*=EcF{!_xo>et`W$vIyLN;Dxy9WfcFiQh4CvFl zJtpoji)V!nJ__9FzyTirxW!VY4EJHNpRIEE2>4(!TD*(C97RBAaYUyHum?bfQ>@R? zagO0s`w56i1d#yBz-5Ia3RG{8%;dvaF0i95vE&{BK8#wqk6Gow8P14fDfwpSvM8!& zh+MChE2OsBt=6iKJR(uzO6f_3!wGxzfkzwxQzIR{8ckk#E%H69Hz<3hTfpAZ%#3HLj_@Dz!fDXdoSA{0*G~9L(RU-n~ z+#_P0S01c^Gq(C6)OOJr;@GEi(sP^EWMdL}%c{*|Dzi=czW7xRL)6HxWmQD&Rr-yq z*4eBji>!Zbx3!Cd6cyF7Vf9+~8#A6*Y0XQ#Iw}IMz5%~^)ixkeBK%YHuIQK6c!b@A zT_wY4i81wC8y7KgN~+Xu_|aB_r_2L$y58)PCxUVu7H7_jc%@Mn?8Gs0ynN%ZV$=|^sz5|h5hHd zK4zDF^sQSw-y3hDnr})&dJj}1C~e>Fz79&<7tX1s0j&uTjQ~yA-Vn!AOOH19(3=%V z)f(n#LD930uoYlrb2lO`a>yC85@%L(9(mL9+K>`wcuaG3UiXyE#EV5}_I{bu9q?tn zlyR;ad0(iMG*v`v5H1a~T4UI}9h0O`PT*IIg{?Lwj`1I;>5Ah(w>!Cr-piv?jq3lPkk! zy(a>&X>A0<>5WL7dsNR67^)Hxj!liUtZ;YB-b7$Xo9##sWwnONfF*syAACU#!7Br* zRp=w9V$z*ddmzGD+xNAAcEnT|QUi{)0yLjA*U^UeP2_ku)-tdMuII(Ap zW|e}brxmsX5x5vagd6a*y#PlAKto|WRVYYt^k$m4-0ATY>`lh4qqR)hq(Ca*D@7fQaEfHo@W^Lh~+x zBu2dO+|GG+&+WG>+!pI9x#DVR53PGe6D}j%L}L?qeaSgoTaBv>4BWYHCO82;rA1-! z_MrsGp;-@NJvGR>a#)m)e(?*eXAqZ`##(D>ZL+&=zSW{=#gh}{5W|(Os-~9Q%>yto z*V#Cl`RzM)SY1&=Y-U|}poVG%`%p=T&9zgS3ls3;v^2CpIIRv+X%(OUO>C1lmp zM*l#M`Qu75AGIj?jC;3{8@T`|CgvEAZmb5l9N~6dizRw2*KZRDMfnjFD(L~MgY2yG z_;!2nf%RBXLJ*YgDX35V`*=FOmv{%JLn(vg1AmIui3Zz+RqZ`5Ki-Zw=}de77q59- zBRe*&b-;Me;v*h+?76rjj>2%+t6l>ue2s*2#-4CBnpC(5WkqKB|oTjE%mI& zxinTMl)w?fbV4XibF4WTl&)!!OHQ+*G)Y%ob(PB>EI{yuFMPqSz4ltW=bn4qoyv7u zWLF;G&`tqZ#jjp)!3ECBA*SfBeB~>D*T|fF^2v7IdFRt5&NHUY5rzW2SZ&-*;#IefqafgUK-t?0X+ljK%?*GJDrTNRCoz@X$!$;zrH&?v=@ z<^##Ds-u0{dX`qXUVCz_#FFt`lWi>C+`GVr`0!s zRefBFOzKP;yr#A`32t?+#zFm2f4zGW!?p*!_1(KSwe9s^=L$4SLEy>z>W|vcJ?R`z z0WkH--u+J9SN{IDN8q9hE_Q!wswM01<)8kNef8V_ZCC!_JJvC))2{fhuiDZ@OZRBx zW&iNjfJZ(Fu#LULd{b?ghkJ zte3Qkwj79l^aayPr`(PK0cu!5Rli8TMUvv;iUq(UOcb~Ri@<3ZTK!6ZYfTJ53SjR~ z;+~B5JJNu~lLfy%4!A?hKavPp{}$YPS*T06AJ6v9BCQi~&-`wT6OX%t^8ny!2EfIA z0kp%gkfxCi!dO{57anf2NXZc=cFNCZCj!X&2ew)Sui(IfMjP)U&oOby69^hX+^=&3 zIZMJgS&7S>1dr(Uq;1>O4}lf;LwuR;s|t6%4D6!?fOtrP2a{J-3F|CO-4QhIBKs;K zxXC3?mSatBZt_ELQmQBkGX;d0I1`zBW<1wqHx_OBk%Zaw?wjtg&ZV=r@Af`tMGxUJpW8A>Y)#x^^Xec4+YWgt&z^opW3@I+|57kvx6hhu zJp;QC@Uj*I*jBfY3kDdj|mQSU{RZm&OE&HezVgfE{+m@mQc& zT0EApk*$4G{+yqIuL>xfW~B_0_wRv95!AKqk$tjHiXjfnNTR=UQ9f5Q86;SJ}y8Np(E`-8bN?C>N1PHqk=+~N@MzN}x z=L*)qcrCmcSdzyf$P{OWNF7UKD8j+FrHN`*`-(}SpL>;sSuxv^9f(Sz-zY|hS>s7v z7zEt=2>p>~*>o(2A3jNl_m;Q3g?!TA*y9#2zx;A%1n_nW1bcO4X_jmAx#ynycLQUu zeZ|^+{No=txoj(ilw1o}u3YI3i=;_$$SPm`>Q`+RInLETx#C`Z_0?|wboxhtPVcS2 zt^T%E1P1k1HI~lt|bue<-Yd%Gxa>3qmOrO$?7Uz3MYYe z8KzVuqpMBTQCpq@K*#i3{q_K~H*c!{np1S1EY-S}+R?dXP^^(V@Y2+{XFRc#pC5Ve z$Lxnc`Cpf^Y58G~SJF2%HXrzfrtuk44l36Ey6dj9fBeUPe8M{ome58EXgSwi{-Aw; zD5ar@5GSkrz5H~4{CW}e$F9Pw<+gfB!=(4g6IUd+V7&waYtoMCg;9*KUKG8U+G?NL z(F@{SaAkdELDW8dN^&_y_*A+ly-aFPz&(d>Ah+yFET@_XWPR54^a8u@E|mlm2Eb6R zZXp(28FhyTM1YdPvP@huI^%gZmtfa0jzIy!CB#?1g#Vz7t@UO*1y6zQL$6N8rQ9Rpm)ZN-8mxKBsy zVbXC#)^yp}Y$g@rXDbjy2L62cS)$O$EKJC?Gb$QUUnv+h^G zU#!*w{1VI*ca}lZJ~_3bs0Z+Cu9H$iU#zj>IzN52qn;O0{{lC zVNRPBYQnZ*-6Jq*BWMwi#?!IH{AMcw&PPyE1R8Nkhb1*V)MpW5ixY`j8|@w9dbs_8 z!Am^T{!|YEawt-wawQtXN*;#-HBanvue;9>F?8*1f(k6%o?s=9#HH->G-tcbPEthM%9=48TIDp|+ICQ<-I+^$Kz)ZYbfL>Tb>=qUPF z5Av|9vofO`RR~JIEnIxGHNnwfxJCCt1UWID1QbIo382leresh$JBDf25s`xh12t9~ z$XjPDZ_5@m*k3mHvc90$MiCky8-Xn3x@#$#OSpgsn1>bqROELKfSewIpMpEBaYTJ{KWmHnp zKJVIUPi?8Fp7zV*V9}yQ4qU4}9ar8~#X0M5N9X9{m8rl~*VIRS@hrZQxn&jBKK&i& zT-~4UNq@_p0ZYFHzBN{=r@p9YtNQA<>U#CnzCf~mdv(1&mA6kk0((ojqpoTGV;?iq=eb^#9#6U4_X`-y?MnbG2-%Z>seo||ce((?Ks&!d7>q*zWc?vW z$!|ME-rxW)ia=X|KYqPLqiClU_Zs03ULaVYOfQhWr?Zea#j2fESOY4{vVfbc$a2Sa zFNmZGigGE7_)CCWz*^8-FI)lXLi^kc4Cs|xwlq(2Vbcq!7fuuS1i)9Rp?n;dFd++g zDTelsd&xpFQh|0m1eg?&R+z~k#l%yIi*o}E)G1tp*KOL!eIW#-h}i(-$I!;7A(;v= zE~!zGQb~j6vw>dn_3~|uiL(h(FK)tmyLVbmF=;KNV5-RCs*HeeA~gPmd2{Uk0ayVy z4qLD}N~|yz2KpxpZx*%sIA9xk6^r6tJc@9lcwNOcj{~&*FoL5dVDcot@9-GfcLWzE z8>NY=#$9=cu?bekEx=@>u~E^O>)B|0G;ulQ1HxDUx3~%1*9D3-R#~hMfuvmb3O3TW z#Rgh}RzI_udOjNkm};W#d1RBdwRc!$19>n9Q;bKI_3iAknjmp9%z-&Stk_Apyc3Vj zGq1!Y-C08zk5T&6LayWy6a~n3PEnDIv7n^JElW(9*v|c9C+_rGG7e)o6w#cS`y6?HPc z?u>u2WN^C^H3$Gk8vw5OCF7f%{O96nk0^SZ9tqdYSc&uLMN1CSwuK9=H0{r147zU!GuF7?{gR>yTrZ40>S zqx}MjvhH^990Y20T<3c7oPJA=)8B^9QT@ioMn~qR_+J%WPk>W>aq=?PQHmxB{!(U$ zAzMZFppUNWeY9VHhw7WItK#)l?dke@f7KrWc=ge{p7RFfxK~eYsvT`TFt5I=-&29V zuA_S8^3;rg)`lR@uw3?U1TMVrLi^AE{7+l6X3Y~S9D_avHp9ZtddKp`0?akWSyW@y zQ5%OmO2WY~FQ@*frA^91umS9J?8SQ7;))xFeNv2Q(je(YbuS_du3ioS;UM4jLI~)} zjandCf`=oYVqw=ly@=k6rQ;O{0zr}^>bwB1(|Tz$xNX%(>vvBux{fAyy`U;AuK;8~ zH+4mX!kP4Q+^Xwgh3-${Le#=OUQAhhX-eYh4z~$CBW0o79_k8a0 z0^x1fUv0m+=9^Q8ul#+QjDT34Cxu-Jv?yA^bRt}ISs9p9mDg7oI3%Ddpv?MKBy>et z6!L&(Sy%m7ode+sYnxqZ3l6Wh+3g9N(d@UnC^_tXBZMNcMotuMco2SzV2qj52z!aJ z4wLc)W#f1uX??j#S7wny25>i)J9ZG$s>r*7l_-r_H^xT?35QXfpW=zNZxR%19hhJZ z&&z_#`pI(W){zS4;sWu!vDyy%;${DBKe_gLyZ@eh39spL9%EYL$FT%!ZSkYLo`6?^ z<-#eHS*GbH*%<)cYTrMBCzL{vS)+r*MHdwZt-k>VtjKkAO_ocy&X;su06M!|(fEe2 zqHAo`ZYjfB&3H`LA}ik1^6q!P+gbX&eV1N(sRN|mZ*8@HFIlq0Hf-21b@!9{e)idC z+vd%iUEK0rcirXMPKB7`<@D`wQJ~E^p9jF{T)ue z>ysVRcL6(XJ$Y2$C4*|ez^T4_pi##pS2@{=_9TaT?dth?R$;kLOHS2q0Zjp0DLj-) zMs4UE$Z5b@QPDNkhPVk%J@r)k#V>xb8}zEr zQ$MP!_B?Q_e(4@1zpFhRS08k3wW+@gU0285{PHMxRz0<=;rq$+HiqAS$y@M^vZr- z9BJoOQP$D|KsL(c?6mCcMjz)E)k=Wp1kA@JK%HFLhEsq^iWX^U*-NCv&{oIvqVsyY zbWK5GuH@!BqhkiwXCY^y@9%>Ry0T<*4x3A#L$+(}!&X5&S8fEbimTUF7LU4_tyUK# zeika4|z%e3xRU?%AGhyN{S53Z_8TS*&oFXxYv@nX3IJY&a_2ynr;2< z_gK>@N89La_aR{9t(tKgzv>YG!i3Ev_NPOkgN1p z0@lbE+?`nsvA`9?3wI4O_Kb^w?l@Lt5qZhVoW<4L2{b&*2uLWJDMDeLR&ws=jwr4Rspe5!D> zYiVy~Vn+BPb|%3yQp|Lgw(2^YYexnvjB1FLj&FBt3}}wX+*Ld zAzrQyGBGAw5wBqQZo;)2$AxlvhDYE~2Fbs?>Iw&tU-#}$+2W&5vg+D;_cSI5SJ=4v ze*4KcKWq2gjAuc)Ov@3F$FG2}TbtNQ*(yLMK%A9;EYBZiX|YsmjT1qaoUO$^Eub6< zK~h*-viXZ@ZO)=8BkjT!&a=VJVh%kIS;|3Ml9CoRT+JNt_@OgysFB-`qz= zYcJ$pBFv1DIiS9}-QM-4_u2LrY_@M)^Ix`Z-5MK%@RH>mFDtyFNJ}Us0Rtf^*8$&2 zg)1Of|3Q=}63QZd)O~C1S6%(5`XZ{-)}>Wo%|Jiverf@8!Zb<*)%UYS*G!@AG%K=9 zsJ!~suXYyVEnBwO8E2efpZe6NcHil`b?fYlU;N_k-`ccE4-~5*S&kKNcg{KII4h$7kcupiUh%+sKl>(XqqVC&R2_Yn z>?aVam}EU0fnb45o$rBS)e%ru(Y1B1KI*%y(evleH~r1&n0H<8esvA?TfcRVS4a2g zwV`7o*iwIW-)ckWsE&^59Q9vY9rwmb-&J&;iu$SV0>1rQavdAR& z-Tx;P7O6cB$)k4V>V1>!?NRv<`|;Y>zSf1XJRvZS;+iYQO9=vU?+vneW^jX+fRJXA ztck`*tAOhi3!OxQJfJ9m+p&rh88{C}iq<)cdoS#KXtqJPfK3*9z_MPEFrZ1+K)Gyd z;S@l2F3uu%VnIc$aGGr;vo5eQ2Y9iSGDj}V0^@qod{`hIJ0@-FMRXGh2v6?WBK^`n zH$ib;O>GS>(6})J3j4_q98Qp*n0rgKG~xO^Vri_gd9+`%=gqcl+qM8~nRte>)`}Yn z5RD0F6iq(yzw`wI0C|(SZ-k67CaU_DW^yB=>0?5uZECd?X^(0Fn>Dow+|9dTl04Ru zJv*!&>-2>$f0_N?BO7hqishEP@h>*Qb?OizGQ{KhN2KXP`v|ZNGoESsSD|rZq6-l- zEugCjQTL?OR$>Cj5P_Q;+C{)-63e#oOa>&yNG3y|kp(`F+ncP+DvB3Ya_}Om`7tIZ z<#km&xqz{RF$bpk-GHSx&odwB>$j1@pgsTmlkLy zu*xE%sX1X& zzx!wx(U&38Sm-*`=N~t0D7o`eqS8^87cDld*IKg$=h;UJKtI3p^xGGffDQPVhBC9( z$Xjl?#ic)5zka;~8M5Yi07Bs6jcw#2RmsMBbSm0Ohv}du^Zz(Xe z73lQZR$skq>U>$7^<1}Z-Rka3_pEkw9rZy)^|VjNv`=lQJ?+!A)xM5-)^K+c|ERyJ ztD>#OZE{Zhs})`CX`ep2U+=p+sr=7=kH7~ndG9{!e*TKj?(k6sO7rDIlu+!Xot0AG%gT);mih zNmSRzEwdfMCBSu<_A-E>khIAN5KKF+50l>&v!(pDO8tQQGXOSAP9H6v5<0{XJuw7u z@$_3Sv0S!Y+bq~F-hzIHd7(obZ>(vsd?VEQFlSc*0&@X` z2V%Tt;4Zs;%T~9%S0#w6W}+$dr>qvnYFTH4w2{oF0Bewxgy=u_Tv>}igK&|+?Kp?! zcU(eAZ;!RL)H5jouyK{H!&SP3n{{1v*jCJHVJzZy)0!O??qSiP%~Aumh=b!GpEUpu?yjHr6QpQTkhnyxZ6cghB*enw*gO4>hVZZewFjLdg?X5DQE85RkQb`I4v@-~8J21%_)>@N(pT0+ zEV>n(n3UBSZU_O@(afN2+Oo>Btz*_o++rk{u+vtYX$u$3wM#GiBnt7(gc2$JA_{Ra zT8lHpDZ1qjiISW3s=&IWHND8XpkWX{iG%@pJn27mz@?_*{&j;=Y&-8V-2wxrIAnM` zT{=KNmh%84T+^|LXJ`Z@VyeEK^pPrZWq!jO-e50$;R{W!&jKDQM;&$4 zUUMP4gZuH1K+OXWJYaI)R>~$_Pax(+FM5%)6sw8LF1yS`!lj~kQm=R1(sz3VZ+Nz_io}E5YZs?j$9z-S-NcDiK&eccP(D~Z0b_JT%pAUZUgLdbgciL}$ z^BV`mJ&-LKS@-DOi~6RI+R|9)T53z}st?|A)mPDP71i}V-afCc`s01qd1}u)-<@3k z=Q%L~>#jF@$?f)_tKZdcwoV*LM`z&Hxqsu+$(28T#@6vnbgtS1Xzd}fP{ z_lyHnm+<=3SXN#j|`wEVA$B?G`p1Grd$t$Q=6go0UXncqe%ERz9i&Iqn+D<4@0!4!FOS(J)s z!|PyEE+V+}Y~N)Gh{MF_tO;x!5En!s7*Q-&s?ug+q4u-jMN3Ok_80JA-sj+gC?qvc?uqK{h1ky74FwQ+? zv5=F=h&IshBarBxy$LUXn>60m7Nt*FED4GqE@E|-VnNpaEEAXjX*!pN#ai*duvard zYLeAf0uIeFW{564i@N$ci;{||XBd|r(5l?pMtS}%ZS|HNDcF$pTU8_EU=V<{R5#&@ z-eX;xx}BA{j&w~$axTldE%4sn*$EL@AJ2u!U&0#W>?d|wn$*EA@*MN5b#Vc6I!SCZ z7IH{rM*AVU0vreVn-ZgViNBL_dHP4-yMW?L@4A;1YysQyrhh>&o)paX)syK8l217E zd^`Sm=h;2C{K|gwqi@*#cl?g|XHpp0*I1PIPooi#I}Jr}SVeXfrHGyB%d!<+odsSXT;!=47&PMZj$#?uanT z>8>7x+yZNY#jU1=X)-G}N8*r~@1v}%@_^Dbn&*72zeUzuxbdKWfj*gDqb)M`Oxl!MG>#a6#-aPm5+L8Puc42{Luc(i5v6iJ!7Dp8?rIS}@x@`#kWcx;iEqQ!&7DS5~~Ro~x5Pk+VG4VVx&X=>br$9o5r2 zE!j_UnXaumUfUipmsMYVkwss1RA0aKQM=lDz}4HQehL_?kDZ;J?mA*WKl$; z&?*2N>%S3{4LFCYl;-tY^StfGnlhaT+H9QWo?yFI5Z+@Hkp!UtNur*eVr~ z7Im>!`_XLIV>OlLOACNr7+FA-$5+;8xojk6)VbraXpN04*4D)><6ng|A|fZyFL+V? z1MC6tqL0Slx|a{|t_e&Nj?Y;pC#}aB0b3uS9dI3s2CQfE4&0?%EfZ!z#%ahIGcs2r=V-T}w(u%+wo*BV9%a~|_!vxQ`(hdTW4<&-!w3Sjx zMT9Ae@pc4VjGb~-1{i<2oBN}T4bNt9un!9bL@<$(^>6P&F%h;FSX5*48*OOUu=Q^0 zgP^I#M#z;Jt0Uc2Prp@?t5~sZ6>;)6@7P8>Taj_7N8lT=G5V-y)H?{cMxaxlWAr~e z#x?1aa}7tROTZ33<WGk+d-5?xR2D^7N0u5YI>DwDz_q1jU<^DeEL2-79+o$(wGy z#bzFJjLm()*u4geX*m0-7$OMvmc5boXUh^&c-F5$CL&Vyb%b_v?65{j` ztn-RyE(yjG{Jed92QY?6KbIY2Z5pez9^NG>htnt*mHSt7aTel2CTXn4oKO8U3fN!> zB{TH`@~&a7M(njP0K1MgCjuFj_$?w4DU$)04*0U{NH|8}DnV3hpM*^p=S&?R>!9l~ z+oDiB`;_}QLdcSz{Ph90(V9jGzDN+KvPVbc7GksL_+pDkd0<{9eioKQ>fPQZ=O;;?bp6?+5Zt(IDe77 z{z036MZ+=J z^Vk*(fFu^?iWvI0)$RrG8<*F3!>Go2mA~N0c^w> zNQ#vx*@L)|IBDz3%QZH73vmQ+gl6lsTu>x$C$IQfWM<{g7I z7uR#F#fcbs=oo|Ane(u?_t@aBE?WqJ)L}WvG=%8lGhUT1&BgtHjew$;I+P0nWiHu5}wS>k;>1GEt~bIyE#@lNOB=VL<6 z5qDjcK*51xp$gY?Uy|H_Sc%7ITh`$Map+N$E(Pu$aAUu{K_xdt-t(% zjkvj}*%Wj9!k1rc7ry6HcELL@v3qX%rL!Xc<&VF2bIo(<-pYO5&k?{)*cHuhZZ$z! z>2_dAk~Va&1s7>3__ysCgB+4rI;j4Eq;8IsLRZF?BI%6OUXC?YAwzk9bOEp|WoD@a zxg)VVI0QK_8!)f~(n{x)uaZ@F()ywa-wCFy3!X46K~uaEZqBEpKD=&3*d>~YqEnPk zdR0ZL8yql>#1f>UgU|%H&&fPLWO@3%GU*0at7 z`Ux8nh}XT8V3a1`o?DZhn>6{6HEN{-g0K|P0jy2T>7ZRan^1l;a2-5&4!^l)^+GuLvA}`O9B+Zq>3T zPmMeF>O9xCdXK!~-YI_TJ@Y`F_U(S}9<8Wx(?^zK5t<3ys*c>DJqxbnQx$1|M*8&PkmJ#887rvJ8DDwyuN$apSsnx_3`R>{n9?K4X=N?N5$8B zByVb`T>kb5$b<6V-@fno;=g^i*@dtE7i*~hYxu?!)jyPuzwBi%vmgKX$9B<07wz@7 zgCpdwjpA}u1Jkm=*bw;;D*#P-n2keNPe)}{1PpZ(lM7z81X|(&z>MNpUCb}byokE%=yFijYnz-?8^1ip(pH>+=}7l+(Bl@G84>PI4vIOR1N70-;0z zPVU{#N{SoVDDJ;qQ#<-CT%UmWiR9?GNz3}Gv_nei6ym%R{ZwiW#a@fJYC;nQ z6GoyIAWdBnm^1?-2Z=+BV-=m*){6DG*JHuEpHV7Av0!sw6Fe^_UMVGz^QbexF|5dp zQwD3l$kk*i@X=SrLg#VS3t~x-Hgm z^z+`m`4Ea`=^(=_MD{C@S zI0|5n<4Ud2lnhGHFzZE5al|M+6=vfJoO#aEcegHbt}M%PXVSlZ+7(Eauw8_ILcAfB zD0}0C@x*KEXa}L3H6$2Tyd>*5*&FD%TVJYJL*DhyciM+P@*&m%JfOI)(njf?WQkVV zD9|z2rB8l3%vfp8R|r(N1ZgDq=GLlapgNX7TkvzJ*T@x@n6r~2*Y z3bvj_+PkLO(s6yJj*EWl9&~+uRJ`9#ex*MGl*_on~g>T;ef?aq*5oNDzYE4`Crfnk2b0_1Ei7oHiVLqOdgc{7wOspZRnTZ!ccfU1~&sYF_=T2PE z7qnYyXw2&3Gi->MZLx}q5GIY_7s3jx+}G;2&aZE4!QBrxd$~{(lkR8iQ;>N1dDdg3 zaf)EQ&JYt_#j_p-gb_?_Q7jEbP`2FJ%jM}Gf$JVwZ(o0KmAmkbTeq1$`zmM7p0%$- z&*}XwGv`>}u5C;klLE=I0*&LoFz)6F*{%`n3; zH%FS9ZNUYv-Q#`_r5`@#FclxxRtNl;pAq?(9w!yRB7@|sY6$B(94T9}0HQD{qO@BB z-RU|@w}QK6NJ{ayaZpY$ORO( zo^UEb*opWwEGFoH!ftdXUP4N}Sn;TkT|v6NhQ?-;tZ~+s1{Bk^0As~0vjSo9b$!qp z%lgVfL0s0~Ua`*otn(pADua1K zh-kA!O2Qe90XzJp4$?jD1R%4Xffq6;OXDbHOUM>ktkETejR~zk<#JF)zym6KYh1?! z)=oOI?YFMOxqW z(e>Vm;!BHF>{EI(Jub{tCkrDptSHE&029ND#unLf< zKj<)EI7MFGQV)4+xA$0xnB0nv1TMQ+@US5JLDkVPc3EPh^3>N0ByF=xGsN$ba`+Nk z9E@ReC4M%7l~!x}I3T!!MYONG$7aoLXCI^^#WCw28?=_0GXbqwLIKvr-XRMhnDK@> zAYR197@|(nMJNuL>+ur$ar+_?+6fTy6Ym*D>%n}cm$(8!z?*!Ct{~2W*jLgcIW0WG zNd^YxDue<7vmE`Au+rMtYP~yq0mtM8CLS1FA3_NenZy;^l_p=YRQeravMC`r)M0UL zz!gRVyKC1j$QKqmx1`yv&9-yXDlD+=#9!yEaYhTIR1>7cfTW6+tFe&xZ+XaOHMHAX z|Mivjo&WwNt|cf9Am)-L-iILJ=N<#NppECoh|i5XHqi{0;s#i<2T@!|c&Xx9IAMs% zO0i=ZNW_$`swlyh=b!1Fhh$95yW*2R&wSisaC?-BE$KOE-r9l|2XG*T;faC~3H;2$B z5U;rMp`k$wAmqt{EoRi}NHy-@Soi_MzBqAjOq@j+s~cdwPLe`P$`wSe$Nr6$e?JW) z@J|cpJ7BEkFTsKu=2-{~OfB-j5Oe!T07^JZz2c!{5@_pO?>># zKmT95;fG(ft?O6saqPMJBhT~fUGwnn?pJqqnjONxIOf12lRry@*r288Ej|i*!&rf} zZfKohGIsyF1&CkTsq7At1;DK>{0^}$>U^Ip z&CHo{OD-@em*9&C;DP1wM<{;KBbpRQmv@d{Q)PEhhk(y+t5rbCO95BS`;odPwg}72 z!O^N}l#;Cbgar9ndqi+4(5t>_D>$Tnp?$S;&U&%kckfEO@upkx^dat0L!h~VxM6vW z<*;n$S%2%AB6iG4^R2HRMe8tWtD56>)G^Df35M>5I4;^F{SZCY+1PesjFkxk#dJTU zLdp#tjS;2;dv>`ToDq=AtYZF2!i4|;KmbWZK~zikrzm~YM?d<}{i*t#uHykT{mtvU z_ggnL`F`*H_5?8HhOSggQ_qzQs^9tur0Tbdwvx*PB6Uo%n?R$!ONRF1fd!Dg^S!Ok z*Lhy?K&#rjSCcssMcwH=t^|&&Vsi<$D*Xm|K{=p(z zm4gk8R7jy3KQXyjtBFeu;#wQwlLv%_Sh%x;Fl6)cI&vzm9zYGRP@;F_-5U%^7jh5dd_o5>q&fdK%Z^N{(Un#7&&K{lvm@ zJz1b-HOe8lNRO6da?-d|A%twb^BxNyeI&86aho-x$p(6`)W)i8D3!L^bC1FWI%SL6 zVERstW9h{yyOgvVNUh`sUP4fsQx~?kpL>>l?JM6kOTuIbxe9(|2pF;t0M1hYNi5FA zGypWqBv;|L;Y`wkRWqy@A*B_+Q~iBbjAI3+%^=UrrEbcKgo|iuln2oY@(bh2mLZi= z0Y>R~w1HT%qyxhR2x6GNopK>ajB@~M@vetA(FewybXz<@>+Rn~%yro2zvu|t_J0tt zq?rH#z}XS3;4`Xh=SHl~SR69sGxjmgquDV_5Cg4y)?}H+wJy)|38R(N-^wViHgm9NySqM>RLug-aLj)s5Qg(eJS& ze|Obift0s@x@_O^sc${*R<><;*ja~f`uSDN8BAp5GHpkI8PN%FEECLCRq$f4?m!lS zUeSuFNei%qM}_k2793fRSSl#2AuK^YOvf&_FLPN@-bb$J80T{yA|(J5CTJh?CH18b z455^)0Lb#cTjP*Du^MX~ldvk6YI*aZq(l&vmy{Dw1xy_?^`wBzV#qZn1>O)hP{I7` zgjib4(N=(c)hKFx5M>6PU_B{d2MQ=#)sY}pBF^{jP2^deW*7g<#kP9&TI<>cX8_({ zIjq=P%j8+*FNK|#Qx{~6`CzXdw!GCAHiblH=v_}+%_BOd_VjT|FF>tV-)qCGJGD=`uFh8~>f)&3Js zIKhpHtl+w~6bZU7fpM>|uCekzPx%Ox!SI0%={e_|vj?}r)U$)bfX}$%P0@;vjN)(I zT5I9q0x_nISRJwKg(gANI4OYS`dWe8R1!Dg5n@)wn3?mV!A7fHVs%eaM{&koYc>$a zM64{pCj?p7+|D_+5n>1D-4B$JkKHkyaJyM>GL| z9U8aN0Q-n-E}`iVd$U}c!)V<=ewxMH3mE_)D|avQ%cCyWd}ER*0FFo5uPIN&T2)xG zQ`}btam9Ir0KLFf#2#xha&Bi#0wTK#ND`vI1%6A&5V@B?cm-{aj=<#I(rP_jySN7c zAy&{4@SJiNHzguAO#VtgpcNyuHP_TJIV2e?;=K?gM#an>Yq4q+3Mq0x5+Lm+EwOuN ziF2-RXmYuLeZi=WU@^$#iD|)*?k{x#u-hCXXeulz6hY=T2XPY){wm6D_Ql;NoI%77xtsg5g{YOYbl5&a)Syv(=^jer6MUn4iiG-dZ?9m9({x* z@B;Z;MXc4mDAA^Skz(xb8-HQf|KKaO^6p!v`|XzR@6i$Xwmi8Pa$6(cC{NJ5ILZzUv=ZU?+v`$pd_M6({%M2jzCty4%Vfnai-AVUfks zERwDeuG7eyllwTXT=cr8`r~aU+t$CF%zo7o=ZAi)o-9;;EXN9SkqflKl5|{K)fZpH zWEZ#}clWxl$yVhO9&evxS6uo-n6zbS5?JT@+EPa`(P&vYD+lLCm~~%XT6mMu2Z8(w z+Ljj)XK)RKc-B&-mCCYKD*v)BCq+Rd-aL81^1laV1p4~=>=(cIh3zfI_|#KRHGS^5 z;|}}gH@|sMuK#RnRKU;!YI={nqW5Jg@bmUht>b<7z@_&tdi$qt^<5v;@jhx(`*gl{ zT;ILl0?A(S&K2O*cNNw1`Y6Dw;(=m$Qi|kC@yi0p9=VryuU;RtRnfMiqr=|uj(0eg zasgRge``|6u-jM5?r)=p(SBs&Jo`ljIx+v(aUyaZ2RUd);4P<>3NbCAcaIT zU}m5%X;rxYHib!b1Xc7%A1Ql?7fR;I#RfnMpzV^S_VA^LVL`05JDF&u5#G9Sqs^W_ z(>AQ#h{dtR)<3-7qBG-w>nyRxan?(~8Vj}DuA>O%;{qmKBkpC_!mXD~gd-JNiCJ)& zLh$m!`U`kTL*yiv-Z6-sDzLOFHcD*S%0V2!wcC&Tx5N_^|5lIdSQwV)!LgLHPUzq)Cg1qKA%aX}}J ztdev{{o|M_s_Lvdi9(^J+YVdkw>Mw-0z37L(=41g&qgy}ARc^!MeEMOdfZ6d8EuG_ zdvK9m`q>ZIPk*_}vH?K9j~Fu0dXCgl>HG-R2qkUD>JW(mz!6HEIKt!@%%JEPwTc3< z^A&2uIsiq84HrTjT*ALTZT_p_GqiN#CNh`%O z96{+&-_h>kzmen@GuYw4%9+>p% zY2RbNdq8vQes7<5zP7raK&OhXr;jYI-gkjV)zfdeM0-Utum^^x_F3oZUIc>WKCP|J z6JQl6_JFZ>uG;hN%ey!2JNDRPO*yQ8``h0-Kg?8#}dIE>CU*%3%1(joaSNI_s>n?8X~!d^|AT z(!YrmIWZ}2an(iB3vhuT**ZIBqo&SV=j=9Kwo2Qzb+gT`Zvupo#tHY^EqAR$>m1@` z;YEgNTGYr908Rn-zrHy?jZuQ&q5Bj}en3_plrQN+eU`Sap4poRmAE}(0H2M%GI!g&xr1_tG@kO zYmTk9L_-?@_(;YyhT;4X;YCeZ9Hl@4OLj#dYNWd$Ez1{fJE1y;Oclvp+v03w%+Etf-Xs`Xspj5j4uKVmS zJ(F(UlTYu_s*gTP{n}%{J52Zm3Xm;Zwmw=JIte^E*8I=^?EbAU`x0jMwtsK?_Gqm9 zqdWrT5qMfhKsl7RZQHibYs&ifxzByhzVn^$*atrF0T*|=&&t#2faWTJRF9-eR#5>+ zflXOW1x~%K_IUu-o8xrc`>tV;45OlbI;P*=NBeaxePrF$R>xFV?Rs0Cr=os%U|7dh z-#g~*7XZ~q?Fh6A6ze?6%#y{l-@A^!t9ZY?K5FY7Q=a1oAAHcUW^2FhUB~oMyV@rL zEh$pwBkwBhl=361OTKJoV8FG~TWh1x}lUs4f*N;`V+0wYZj`oh)Fs#BY0KaSiO%#^Deh8l=ciHWNBLtcfY zTMW-iS>?2Ky!;|E5gC=Pn{|1bnJi_e#;t8)px>Gfo9|L64dC{@aPA`e{T=t%N$i{5 zUT+(>Y=Mb0Ws7IxF73mD&W4TV(`mLIr(I+m9 z*^mC`Y+|m7!3yB28hyWIcHD@&^m6O#`J`1Z`Jn|TAxkEj#P&f)8J1k8B}cI*H3P7V-jwD*>#?WNs!N zm`P8mlNjNgCP^&tiRuLTIEfL){jQE&(nZx_97@EY_2D+&fR#7bJz|+!@?g$n+|~}6 zFH5d!NLK=Y{N(nXSo0ZMln1i1%Pl;?eHDpI6Z3U7&unOLoV?Bva(SXy!PU;ikK>X( zsY{*uCtCjgbd12mq~`hr@$()io_bB?CjR(|E36yP^f7>Qdpu_6&7Cu~%Cz|Yk&Dl^ z-W^+>@LpywILumS&2_HUPgt1>a$6Nk@RKRJ&e^Ycqji%aYF}l9aD|V3@9KRWKJfkj z_qzi>r+h|v1j-}utQ-M>ohK@Cg;kN&@K7isPW6fh485(l-`ffh?yc-=|K93O-RE6r z>Q>*qkLr5Irk?lM?^Ek|=L)2%c*nf0cdqw)Z+&01XpvoW%{99LxL3#9>Rf#n&=pXA z>~*|)-hTb|6a?Ne0r_&-zY!>d;r$z@zo*ver=M;sSFUsfbP7Z8N-!2^w7%>vi&PQQ z+E1KNb;yRdKo}K=k$M0#=|l)HM0z3s&LAwq5v-pT5N>Q;yB;l3(2m6_y!Q41_^ZZk ztb$y?SOCNQeo_|=0%VJ}5`aB_(Hwi|j~lQogH*c4e-1uyJ4R!F86OJmLUg^9uKh3$DW zWKLK};|+~4Jr84P#r3*(P_u?hCp4HGv?!V`X~y&sL6-9IdYU-_W#wT8Dsv73F6YGn z1xc*Vjm?c#S6>GZ^w}H;#a69)(25J@+43Wgf@!;llv62dtc4g0Vv$e~%k{{gt)lA@ zz-_bDw_sUL9BFfxG+NJ^lzrpUMRxkhTZq5fisg5i#VhMMNZThvjepz_@^W zdUI2Qb@z1P_6@nqD4I{s$u~yLH>P-Pw5(_@6?z=dGsKO3g7L##ja3^aMidiLpomOH zn#GYJV!gReiRa0mm$P7df0*U%(>nsc!)hqzcUgnqxa=@H?MK%>t^+Tg+i5>tzrnuv z=Le=GFy6Ig{p0R=Z#&NThkv#jyk``Hygx-LpWgbhuk26Neb)I~z*uu-xs*qsJObqr z*xLwPa>*r*9BFUIw6=G2bl7pn9p?@!e)Lc%a_c>miwBUWj)C^c+FUNr+z6Dx@Y6r` zr=511-G2M+_L7&pWcS5`L$=EnoZN27!9nwZ8Hc)IKn9TZ0#=odfkekP1P~NClgmq+ z%2{GwjWtIat$Rzaty`V8x`kENQc-KWHuYPMHpVhoBjZt=uzC^Kq#z*ZH29Hi#*Co0}|Om5QYhw)mmzo+PRD3_KvTdY>CWG zR;p^m4f`Hqf+4vokV|;PXT7CcY~TU2xhKNxH~N4D8jiO#y+v3YZQHfmjk~+My9OsD zxVyW%yCx7waDsbqcWE4gI|TQ}-5b}<^S|Go)lq|*)T+BKJy&>aa3g7jsNJt(8XpL0 z^gQw}DccvsP(9-xyRPAn#eP3pfHEsjjAnG%c{q1{iwa+Wdb5BKFlXrk= z#kwCe=R`WCynFEHb459z?kVm;#j}!n~gHrKMs0syTDfcn%4KpK;FlsQ~b=UDCJLscEoyx z-Tn%uQ@ihR%Qd+S2$USB#vrHrIu3~0Y~9es+nJp&dQeQZt$3#C>7j7XFhkm?%3m_;fc>*{eU)AFxw^Kh-p`>QAXm zG}KF2i>+MN@CcC+N@7~6uC8j{txa!XMPJ3!`AMC8+Z?omwr%YS%2{@5@JXfrXqU%*QQ#RYOU3U5%wv^V-C6>J?qH-PzUXze5I)6=n;j zN-@f|tA?nNVbEdFdfDjWS@PnktSY>6sp?@Q#8y&_>`(lEG8&`um%Uz8n0 zJZfN#9jLhAvBq3=z;4`>@{No*u4^%$6;jFb#<}{B*J55M7_NZBHyb^;=+^lmI7MgC zAvqM9Z@Ya0t8j?HlBlw~c|cEg@v0!0yixHY!+9$BRgrJujJjC})nsq^k4Ze0xa3ZC z3vP@e(5xBh=a|J6uFCnUsL2GWV~TvDe!8`KvRwQHK?P?df*9Qa+@LAEZD3NcAU8r? zW)FABCxum8v$&mkUSV`M!-ym%hkn8Rsiti7s=fhYZ9A6f|2Ir(z;pPjb9b*HHSa*<8K8bNvGc}mU6ELcW6~P`Cf303e*X3rjU}t+&yyf;^78Z3Mr={<$8+!= z`yM7kT+a;wa3F#%aYV#PiYu0&0vGw|1KtX5ojF&{`@gt-^uO>o_nIzaO^#-osWQ zqq)>WEmT*n1*K!^OLp44n_*gE_^FRc-+bYx{~ZF}Y&vOGv+Iu6v8HZb#Sjzac_c=# za27PnUno>C&VsQS8X&7cVG=h49HK@;SqYcihxxvOvdIEanb;oGhsUTH-^oB;o<+d! z*#3cCt1$Lk=o_z}i8cE8{Ag1O=)8_irl1xrPTYNpRr|{ZuS8{y8+n8)_%=bEGwQ^d z3}g>9O7Idu_Js9~p@qyg-yNmN>>E!d{_Xm12}SAFtp z)ihPLLtKb2;Bs#Fdv+bDuI<{$ZzA{oMvF&kDJd@L0;5+ehl6AuTz^CNgbif4D{_Ez z0-??I5o<*a!c1#Jl2^llQ?%P4)3*(;#2?Mo%^FuW!{cfxt*;Y!Hs!U==3>@ep8UP} z>yF=qoNF%;3I(eQ`WbUg>QW%A{Oo4w3Hq$UuAlH8$#ok;g%hcT9OpF^@0`icLAljg z<^y%6YXw(C6eYAu|1c6U{ZT1WBqX%lbbni_rdD{ew&pencSjC}p~mNr`GrqtWA{R_ zgnu>bFlA@oma2LgrR;Ho;c7bl)OzYNfi_4DurN3q(PfzmIbMKU?Qlfi!RzBvd~VTA z9h(({8tTJ)FKfSO;?)Vo$EM~8Rf*W*`PZ2M@@3cO^^HzR3 z7|acv0-EprQbfCJ^|iSF+*KM>Ne6ruvE4NlHi%-q*iw-O^;PWPnyj%7t%$uA_M1N(LFXeWNDnS$dlVoLwlm6WmM9{f0}}= zK{Q$j9Za$lFaVs*UcV#y7&Ty2(Y9gKZWxK;7S*ID4z_{BCF`DyMs-JIh#^Ibq6s42 zbjoU@_@AHMIMyd6v1`!-)#JdSmF(~1_9z@JCL1J~i_2fT_Epoeh8#~~vzy&CyR$n? z!b$6V$QaXp7NKhBXBkW{Zvkg$fVEqIIH)l9BO@GuW{S-~V8B?&K&zJ6A5ks4x*Wq@ zk-HC@1?OhH$fb2$>mkZ7%zAF?B6vU6g|ufKLRa~9BH~M$`@GiB`Bx*lW}MAO z>;`>N@EtKTKetF4Bd@hzo8jRsR=pF50V8<&&0z=Iih;$~c_=wRoork%De4kMOS$hG zCSRjpq9uh(Mc8SS^r6he;A~-BZ#g;52n5DN*AbpVfcdvu#n0;kgPe5CPk8%x!E~sz zkm}Y+BLHmu+aRlECXKGuHs110DG z^7lID8X-w~Gelb4q~(J!bh~M}+|;@#n9{iX8`u3*()rVbgh{5G@4mG7M%~92U6Q=+ zK>|A5>i|Jw9+>KOeix^@KK<;ha6xv36}09sW;NM_pY5CC`8lA;a!2s`@-9KiwqHqf z$D^{!y!mM>S!>ZKW&p?Uq0MV8m3qN#&P>1EirP0Ggw2oC?v8XBit7Re+To~fpxh2l z4*kT-WkH$hOyF(1_Is8Xo^@j;jM=qj-URbb`^i;|Y{!5l0Wc4hOOQg8z%Kx@c0;2F zx^VWqe8Mx>qnLy$HfK90*o@Mg*`tLUf7+u$dBFOX#slRRk+QFGW?{O%Pf=*53Se&a zvQ>AQjch>0f-r5e{b9j3&o1CmZB-ji8Bd2c=Z4YfIWAxEi_+rTVQcM+6?DkErE;h- z1pUREC~YqO+LT=1siQdL*x?_lIb{HcFr{1SM#tG#V}2UsSav4F(ptU}`HFp)8bLwd zM;lzrjWMJGQ+$Kv*h{_*RxQi-glCe;MqjxpiEDqBEjw#z@w ze)z@At2HKO*E}-mik?rAbK}z0)< zxHgda@JXT>*}`UiO=A-)9%&R;D;@)Fjk zS^XjYz}TLRp3Ch64g#Y~Gw#4*cY*yh^V6x|H;9wgn~A3ToY$R0L-UUY$D}1tb+u_` zw*}B?tf@a3Z3X&%0})UGvP8yToeFOHOEj#{mq4A3F+NuiN;q>l_hmC|mBV$InqS05 zHN+autq9>LZl<7Ra$8N|$pApGiBz#;Lurng7DGL+I)Mybzz(q++xDpb|Y0xrR>r4rj6JHXg;7cLY+eNQH;)dXxxwC}O>;TjNkF)*8% zqHQ6u3?7Y+BDfCdO0Uu0S?YdFk-*;7BV`pOb~5$|66Ih#rh86E1aSptiPu$EaZP%j z95gLoMX=Z2*(~9)%HG`u1(dw(PbPu$1j1av`E+1ZjCp&xF61!>mJ=WU1iddxNDhCNUzQz?gGN&Ba1pXz{Yg*2^f7xsHn)k}= zBssYjUpnuu7@92Yb&CSuZa!r-b-ifvHD2*mtkg}YwR{FOH!bz70ErzU^-g;66O7LV z<#&TDKZn@Vl_y$D814KEA}(wFTTV&H+`oEy`2>G9SYd36-*slqAy41pJ);lnQf_rx z-m4j#q`)ah1ph%fSz$YA3*ovmTPg`$Se77hHzwZ{Hb;&BjoJ$%ibqg~Rd;aoO*O}Q z#LiW7DjLIh9C8VXzuxVfoV7}5xi(?>e$qk$9@|SN_=dEBtQ%?^yrkA{ubt{h#b=PR z!*@+RX)XH69B-5aru0bf?QF)m&0yD@+JA6Mu@k)51;O^Q`+62wBj9;_PJe6Id)Ut2 zK9dvt_TwXRsCj>|-e365n0qbVb%*@i4+?TLe%|d~M%BqemqN^Eb=0fx4`6`pi^3z1 zmQPw``7f@>8KACp*^+2~xgK+z%o3Q|HcDyNBKu~2fuO+8@p4Lf6gtbT8-39Msi7(R z>zbC4X>YxSCLds5_f1E+q3u|(a=5>7Th9QtRZ|j(%=}e{S?SY&mMLP@5LOz|=R}+f@4+xOa^giSXnt5Xbq^7{|a~Bgs`y4)T3hhB~{adM_-fl zbZ|s93e#HnQ=jz}Ut0RC8B?{3ZxK|;>8Fu?*q#{x1E=}3%ie^<1CDBdyCSaBitG?Y zwo+_7ck0)j$?~j$C*p`?A{AB>X~UwB-qs8PbF5V8UYG#bNZjt|?uHoX-sJC$4He*4 z<&(6EKKFQi-i?LHUJ{_p+q3kBeU)nNL^*|~cGyn%@V{026Y`EB zL0MSMVRV-u64+jY3R^=nYMp~2!c3hkPLf}`W7l4p03LP`UXxX=xaE{4U9-lWk0}VW zBBT@>w3RQ+Af1C3!e-lbYAs?tCuFJJFqo}9*zrg9OC)s_ zbL(GB-HCoPgX6aK7X}!&EwrCwPY#WHpU1YBNT^em@QoebRwNm)&+D&RdA+tTvD(iU zIsG4ztDE0~SoV`^{mn&E%<_8DaB)q>-|IKpg{O4VlOV0hX|t|4$eCUabwp`>W>Kq- zsYxUZ0Ir>ag)n{J_gDLMcQR6f?F-qF9oN*UZ3*;4B(oxgnAqPJzT*i&q*)(7^I1DI z=Wqr>rew&%Z&cPrk3GhJ^Z z-Y>ZX(vlk*f8`C|6S+hB$?qipg z5R#xcNbr9FCsRxu>CSbI;^-)x?apoL^DxEf7yf-J`x#y*g^OE*^rzgWwAsXaV{q{f z!(3s*0H9l#ll3u)=&r+9VF4V>AjlKz;*`ny={>$jKa6SsbE72=liXZxll6xD0_6qN zD=0LO(*VZKn)irT&g*<)hT<@U4xybNn_=-j!g@?BV8PndWZon_svO%@M2kw#Nb3={ zFnJIasuhE)HS*|J=4nmh=px0GC{F z1%4c4hykxOIlhT_EUCTPJD8QuQ$5Z%Z=H+j1Q;%ax99VrTmJnm;Mp8BkC$26pOKBE zOSR^tt48d3i*6>FYgyrB{43LA&CY74w-dI$fqML@rgO^64ZS4br6@)B)juks!nZ}m zwrrmYW%=eCi#oT`h)zmfbuX$QYdx1tSmcfv-@Dv)DWzn^_=71y_uqX@wmRZ2&&urU zaD7(puqff`xv%W{DW=Geb#eClxwx zS?;*FPpagbaJno>!a0giIr?^-Rs1YFOATztg=LDqvalrK_QP7nVd z;u6r}4(WvQy1A2f7ZU9h@!hWIyg69Z=au{Ae&DgR+)~c}r*avr_LVYeOGF@^YNkn; z5^)u=ti}9*?HRi6rlc7}DuX?>(pHl5ELZXLewq!RmF1cJE;jo-e%Frfwo| zqLp`v@Dw*VU3VYy+&E2U>qEHSzYo1*i*HnS5Rk5UB-ZeA5Kp4aZH*8=CCI4W&hzN` z?~D7Bq%+g{I!p*5I4E!|i~e-dplj=|kjEV3!BIXasFbE5s28=;xDrtZyO!=!n9k(M zzoumy$Fqga|%s~krpk|PC5NW zv1){}zg{7+rbS#2GLFsPzXQ4ujFwR<9TA#+M9GJR+S_S`JhC3GYz@-Z*31jmFVb*56H!bEVk{p$BQw*1VWT53oYYSBaCEozcs{~A~Bwl6Z{yU1EymYbZ8zVd(I{yW%G2pv^#nXS#qPl(ycuo=apL5LTVP3fNHg# zR4Yuh%phE1(I+)YP9H*GuE@|Hbnj4&f{yZlEbbqV2QOFoj8k@xNo%Uh&BD0vtpSCn zfsaECfB>~hW(1i`s`03tLrQYQxo=Fj!V*<+y2>V$e<$+8<2ah28ipq(I#?4o*7g1@pZ6$T9tdC`q*E!Ll8H)w0?gy(rq~x zxU4C@*YFyYkqR}>dS-?yW-Vw?!JQlfrR`R_HWAVweQel&6~L}d8F2z?`d;yR zF+6vb>g>}6nPgwQ&NQGbxmy2~msW>aj$_9s&wZ!Hx2-QBf21cfaUzGwwR*ZNvXV|h zpu>T_l*+1I2y>zRm4LRbdE0+x7aWku#KfE&U6Qa~EHBO%cS`66!l}fC(eVdlD$%Ua zi3lP;pGdtq$|ZNyjtMLssIC77j+69-n9p|Wm0NGGMQg11Sa8>d4fbWx(Jp%sSeyKe zsadL(zB7yEVB78I@Tijco*44b^d(fVtW1wAZGa-@zzq-egHiU!a)_ja%UK2J=VxBK z;u%v_D8z4vWa|Nob&)?otz7@*{LR=lY#?l|$H|d~WK2_adU`?FaJZ@M?9R|mP_W`! zx2es)k8N-0RPba%+J4;--S+Ng2kyyfauy$>nCo@My&;T~+e45JG zAT!uCDW&^Fb+5rw*i_cu_@xfuEksD_Ae3+H0)ln z8OKPWK_8y<=x7;eTUdo0<0tCQ$O*E_IF_XC7LAG_EXUx^L{`<=GQEk#+{Gt{G9sUe zu03T?MU(*fVKYf^a1$8BZuumB0eU}kIyq%y&3a|sTYDzIaDIeniv|c_?Rol*Nq1?; zNsGms6QW#CTf5P9HlJC7_xFO_pB;%zux(IFlLSlIl7H*2U72e8aw2CB^`A^% zETQ%r#wGC$HK6({RBJ_iqofL+w$Ouheq)@kwW==~dK>=2@wq;$1Xm$;tkGSLn>>7- zDFg~E-2Kk%$(MwYb_RFB6QENx#aM*WvX?GO?qTL^^Pm`+lX_AX{8O`Qs5w+OzW9^U zIKM?oXNQQ`l-X7!1(m{7syK>4#2a4H80}KRb8V&{W9w=O~A_~ zvWfhKmqRBfL&eO39VY@8Bq&Vvyb+Ptn?_Aoh18DGkT;-H7x9HGAt%~)l>i5Vpz}Yn z#aQz+b8Hk#9dmb)lDsTZI>J`FLAXu&%5ftEJ4iJfsrMWf?PIPof6mrmQH4gox}idB zio{qJz(W7uUvs6>g(j`LF2udeNmg2abnvZwsO)@qyWTemwPmzPVnnwo7lT_z%Cp0< zQd*ueT>8clz7*A|C&$7d_+5(I=lqITl_@FpPz_lk4=LO@C+f{ejjJ6twTy$jwa;!3 zgYC&j5#KBms5QQ{q#RLK-w-$`wDF)komZ;qsugCw7PZj`kxs)dtqHp$QKg>sX$x0G zCasJ(iQL2#+=ergPw&7kz41f=go|ZT(zq=kjD^>1ea+MnIdseZccYR--PuFV!V|)T zd69Ax7(H71aTJcqj5_**5M`?7m}fe7j+Nm}--Pj{(D5|LRr++}0X>9OK-6TEE`fvCmZ0gzmmzhlGR>O)?M=q3`8K#+$+(r@MPla(C z?%hr;o5{iIagao`Xa4uwLGo7+oNU?Ky?wQI@BvyNG`;hM(A z?@1>U*2ECq!xzZZ;s>+_bnz4ybW;#%`niG69*TVt@#ccaPoilihsPEfrJKYEeNb;^ zW5dX=baG|E&lB9Be#ewMO7vNY+k;TKNESbbp{^uRK%(kp*cMpFxil1c%=H71Fe03% z3y^~*vwnDlzG#y(e!l}<#`hkoS!<~O`_uJC=9^&UKL(IkpRI2IQ-vU$I2`eD`vf7& zWIzY8qu}INSp1WON;epe#pXuCJsi4??=YzB`GWW*Bcf)uF5I2F-<84c1oct0Hr^Yva5WKIaQRi7E!VuJ2iuSPXP zcdeZ=uKB7$=!~$_BcAb(8s)%`7PCMBUtfjl+9NutC((ga_A0;ce=&X~vWS*{TsS?h zj0iq;-c$P4{Ay0WT5&1ksKTibe}(M2gJdH#;*@iem3@{58FR|IJz#{& zlY-P((Pm}i?7m&FF~1pZ+Mn#rd4oo|fwlkcGGaaZ>1x4Iuo#H@SKh||VKP&WY(vJk ziBBxxo?EouTq!s?n*X81_#NgBst_H%Zv7_l{a>B6!Jqf8H%z?9?IMwE&#TqBwPuwP zBOd{$&~7%RwZ~O*^8SR%cD?_CT(w+rdHJZWmp%5AUSxKb546DhRq?Ft>&M>Fbkm{{ zGd6`Cpbn8{J$SXre(|kaPfQ=%XIYU#5U;^6+}LaCDFdjR1=s>&MO4 zMwsG&S`b=cKDI7^qY}b%ur<6b1}FR+|0xxg+JJ1KA0tQ;4IQf>{O6oqbtW(=9T{~t z3pQ*S;eBQGOBOT*vABc8Uj@TN#C%paM9q@Q#fS|0w&2p- z-lR0#LV1DRwG!cbFx?f(LMyB|7oV`g;#JObikHbNBnum%b{G=ytxh)kZ9}(2%oN-gww5jFxx)Qz zyOaNpIZwRo~|d%hw%s(m zMpbemM>6t0;AvI4b?j{MT?J%x|5{?;WN4z;lM%xEtrOeE!C~${2O%7lj~R(z9q{|x zJ@2>MF$elAm)?zbosZ~22m4_RX5%h>QHSd&^{T@3qcHdMI2CM^5#rwAj>7Ihz1~S9 z30*uIFDQgy8%bTpTQ-iV6J^0w_N~h`+wDP3(tOsgIENuKnu&xezWieiLm=Oi7<8EB zt~5Ilu+%_%rg2X(Y>m}?oQvj<1 z<=Z<$hbp7^H@{eBh)i)y){{}z+gaVcVGuro_s0!4JOLV@4U5KL71$%inA2Vv z@I&w0X;rCMG09;N7+Ge`6%>e&Q0KYnhkH0Y`9f$M?^jKFbY%z>7!ex`Nl7$|tsAcU zV4NLS%6+SJ718^)y|Bx9$9n;JTDN;nG$X@2@N~SwUpjHzZpf+(3QmX3EfXl&uPlq@T^Ej zM3JeoD;2aaGq}U=ZG^Rs)6PeSaJD@)MNev2ikIidOxxlA=*m+>np)&uSP4TS%=E*D zp2JDoZ#I%6QUe~8K~tYRooFj_^o3>%;gz- zvhQi&L%=|LDCG}g8L$1r_$Zy4*wxljbf}P3IpIvnw$cIX!uZLt5(Aak!5Hn2@&Cp0DX3G zb@r-eSE;7R>leFfJ^<0X^WUm^m1HHIy6D#j&#$W_F`L@i{QlK^!Z!qe=>IVEteD?p zw88AiBWSU1TDzDO&KVMyQ56v?JAcBJEk-2H(3PYU?`FRzKzS?5qLc~PVM2mU6`$j&UsVuA~TvUFSk zObYFld0gYY%SGy<6~J0q`09Iz>l?*hhd;feTJif?7re(Kx$``IoRAbG-sWOmLE^Qh z`C|HjET~bV!ZKi(T~Z~0t+Q7<=s44#{H`G=R|@D_+Wb~c`ty z6*tBD4MFg8_lEw{`qR+=1Utn8#4f!}t_hGdeqLGh_^2fxiCp+9m(?tB3>rW5;@_-a z1l=Px(M&axpSb-m+3U!n~@$~jiC98KQ()u(j5Zj%CZXTE@^V1Y6&sQCW)iI#M z(7n~w*YXG1T8Grw0oz>9bfF$V>mP#&zfQk8v5He;wjllS_j=dPv)|fB1-Y*|aF8zm zYY^RY{&14D#;O6Gx4letN1RKQtt1s$>sXfHR;lLG%#dhx{V0>LKVQJ#+7^J>4tfro z0(2$4srJP;&D)bjZ{c$1*9*lJuCG&-C0brk*#VrhhFRLb-DM{f64^90rlm_b#rBb0 zA0PhMsuiR9JA~BA%HKivk9@?0beQSltmvmDp)PE}MaIdL#~EfE6ErgHIy}<0X%6HK zh?ojig^S!$If|ds@yV-D8Cw!ipFnyh?*Z&%?Q1RHkUdV-jD?so!iVi;NUX7ygAY@y?aHZNOW4=avg^Y4PnY zIh8JvG!!JKPabs4YO!X2M!lugaEIcR*RWBX zQO_owu(u!xcD5pDb+2M#aE(1fP>J}_F1Y=>LR4D9M`Paoh3G^|Db8A0W55B5EO7G{ zD@CTsTce!dVi5&`%B4*F*M_?k?Zu~Ek>533`oIK^FYW2=O6iA8Wwl+|>?g$0^A$N? z-~cz-_jUu{`Nl0nr6rt_L3#Gs{_c}i@8f)5cZU)qw&YK;sqoC5lFX6`Ol1pph@wPZ;(19{hy?9eF4GOj1w)$DYwb*p_{f@r5RZI1;#QXMxup- z;Ojp@*X?YyQkOg>3AC%sxLDDuy$cT@l_UO!#lJ^Mai^vhMcQi;#JmP5guGM z3JSo(V>$FK+(gqZp?6t)&-1!TbRvzKdDmA_?@~Zbro2T*qY$y?q{vhg77VZKKwIGX ztatY3*q&)_mH&@!ZN8)^AOgrDS=h%pX}a(o$y1o49Fcxx!?G%aY&8<4LsuPeR$R_e z>HE3edQHz)Ri_FWT>S5A_i+lDG{dqT9)bb}fe8^D$r3s=nbo`GWMGne$dJ-qS>86O zoOZA#G%>yAzRsU-syeSLS-mlUd8<^aur$;S&hl@HWTcOg2^V7KiPq1XqV1pRmC0p>Kl#W&z)17&7bENB5=+~7se&XxQjk|;eL2{INpPHF(#lneNF1Y3Q|XkdGM86u4u(xt5cBpqH%|b(NX8+5ifd0 z+~N&J(Az5r;bKZT*Ch*W{^|~V4}P|$&a5uCpLu^f@V|L}Z!h)`%zYinrUk#0Yy>$S zd8_B%Tr9q6$!k}!EpwpVWTle&{uX_kAV#~{SN3nmo(zhKUVKn-wymW6;7V#ftvXA0 zZdH`czsQ4qO3}^Frv8htyJum@f?r(!>v__YYiO4l zuA1_eT*dU3r$k892JgYBZ5a=M?E~`7)eDBruEvXHRGW*EVg^l&DsU)zMm3l<0UaF^n%-BKi&QyjDe?1;@Z+xLON~GJPj4+BvU6#zvXeL zF&#o>VB_Q-2R|e@*767x&Iwe<#nck0$fCRE-0T{3cW@_$SU-^T4u*d$noi{Ga#Awq z6xR={PmDu7BzI;BopScnH?-$Y*9oSXTdchSkWqYa&a&iR-BAA$vm27ACi{b4mRTc( z%CBfwHF+z@2EBC@@nep6r*bG+pX(fMeX>?X%d>V63M*i&D8#(?Kenf(vnT(3$s(sg zeBx{kCDxT;pspRO#?E!(ljM%9M?hsYXj4Jyvq%*)TmFL>vVb$no~e)snuT(7eF}GXfs2MN7Kj=_*r? zN6ADl34;|r3WY9jF#>0!Fr2fzxqYFnPR^FPE&`(8y&pMCZ1#Rxr?QP* z-J74S2ebIbY3D*#0zv-Ol5$bW0rn zmuZIPij(pV^o0S;YX_o`#9b-q6 zB$?B_t?o%v;qk7}&5+Nnhqf8f3B412%%?{15Tog{SIJi|c^znyi$r? zqI)uy_+-c-U0)OIdMsTH*g^`Z)k|uj#hUDY%duFj^vls-8v9?tp?ioHZcQxOzI|6m zTv#cW+4<+jPg>-#sI~lk?sL=7e3rQ>6|UA__m=vaPX=9Gq!}FtMyP9v^ZCVG%cG^k zqmQ_`X=bi})EL#6i47_Dj!g9Cri*_g=j3X z9zp`ukt%7+AMmZ`Tvv29JRBt_J2l?Z4pk1}#b*4k;3$bL6=x?Lyqfg55ySnHy}BZV zC3ywKV@LAW`k?2~8U{hpd|nV(=a}_vX)(C!{cTBKr(gE?M!UO=Ctj1jTY~@PvI3rm z1Ogh=5TDQwDs*~y(?)<~GOb^rY6nVKYBsHO1;s3E_(z$kj!xXKjxyVOzU{roWrcFB z2cLL88=Ak`{7rd+k^ z1D=jW5Hq#lvsNz?X&Ba}@Ffce;Lo?TRlXoVw9{we8K6)?$|p@q<3;LuBQb%ew#6_q zlT(RQ{GHx3pp+x_uU_HMtKe2G+9CT$S= z$i#}%9{~))QPQp;XLPsWlLeQ>)K`IjcH<8&>}&sto#yaX3j9K**)t7i+Wg$W#@iT< zNZ3%uc{+n*w41=VZ&_k4$V2UqWySchEXc~6qcsw^^Ym1Kf0*XEdKq(|(<9ClgCdG^ z`PTwIJ{a=p3iF_KlBFvtITgEwH;=;Yh5^?AJCNE#e341u%TG)i*I}3<;8S+6>ABG< zC=HY^fxO`ZTZ3vu9~Rc?6WEoux|+f|KR-kc=|J}--Pi@6uf29$+$(r9TnawJE<^GE zv+uwz3IU!<1r}S9BM} zy-qpjT-GM)u;ffx{Voy$InUK}zqST?w;!^S&+SW}{~KZxYbtD6@*W5hL%H=le$9P1 zWmF9Aj;bR6ubTHUbTN&+)D_si@v_gK`V@1)m4am7@IOV5c|d7uY8Jr+m&E*Gh4zxO zlnP+r1IpzP0=Vjz!(uY1ET&1#RVNA)1-+E27DPK_rCDqj3Re9aZGy7#QhjA)>3n2W zd+;b0b*EJ)Af69($!4$wQvol5^*bUHN+rC4T(a6aIt6hKFfR*sX&c+uqjp6m zW!beA&;TxefSX@AmP!YztQ&T;dF$y9u6Qi~*rlW%J28w$5)iM!MGMSy6>lN`)yR(E z63d`QGMC9da)MgZrEu{8(!Xh!e%||vM3!|8`DDej12yLk9%0O+8k`a^M>rNSve-f8 z3o}fR{5~?$_>s!X->*k>V6k&75!yBprj+GiIaGYC42Fy=W$;ln+Th_l)Er9cp+aow z8%IM1P=ZwGkPYny7D`P6zufNjC_SGRD0C1Lb_ z;R3nvR=UHe>=hD#OPWLKddbLJlO9CAL`he(;$JnVBD}XvhQ*bfSQKKs@&@%4z7QZY zaoi>nHUz2L!4r0(#+A8?lhv82@H~gK+>ue+kfQqEd9>ZZlMYRuzu!lN4k;(qb^kZ1 zwn>WqZG-h;>jYCX(AbBj@GF>t#aan;ucDePZ8zb&KQhYJ--6ybD`4?gDv;xBqylO_ zB&xLtu6WnE6wtxlzz{9W6~u45#2rq_|K(MPv1~ozr&Vq+u!g-7uu2}R`pyEQ%0u$y z_XF~>(T};ut-LLKw-4i4>#X|&d#9(+=HA_lZckGXN7)>>`1jetK{o4b=!@E#LeKn* zrL^7TZ`=+tqFZYS0>2Z8=4&(F#|K2gCR?m*jieC`Q3l9;_)78E?nbUt3@B--N zd($`Yqu|}J)p8$@q*A*Opqp0+bLwnIM%;$hkfHg!;ddTi+uuHLFi!T#{O-IB0h3Uw zh8B4&{SG53=7QI^fiGM(Py%@1NfG_}R9OPw%nwg|%2V$SXf~ZiomI3TKn8l?9h5Rr zDJ-G2C}GoCPHq0!ZaGX_+D+_!=##Cz_ z;^j+HvDDRpq&(altcH7BtFn5?v(M!^ZxmurQOgpew+(?&T+Y=|feyLBk~;Tq6r=g# zF$RSm2Qs70LWikzJ}iibyjZUu>7+GWc?PPDh~qGYm6$m~#Imc=5zAL=1L{Rz6(Cqo z((BPM6Jb9|k0*S5j1Tv>1G+qZ*L@;cvnpIpecMug7O#^n%b6laOjK}e2Ih3P`#*c! zdGa3*B2Ud4e7`la(?=QWr$=15#^|vAT_F9`E{}dyLEb7cNZ)aJX`Upje3AoxHxS|z zd+H$T08E-|#fVN6G(ckj+C;N6tbf8Z6isSg} zSX{Mo-8HEq8S$OOFB+tnP@0`Oj`So|3f}D-O_#V7Ie!9dbDzvPJx`a)Y#ay#w(y`p zGL5k@&}`7<)%Ke29HO#nE)|BVF&^-julvd0>rncc0e&aR4qystgfce9!SqsXTWJGN zX}s2)6{--2Y6MT@FGmGf4C>Jgd^)4q zagWK0bx&2pt|y=Cx@gCwZf{SE_Aek`y(K-_GSzrn_E&rOo2o>J3n#zOj6KA6(MlR= z;wl?pqt6Llp%70fJ?Sn4z3lFjiu~Z^_{NnTyk!S#21?AD5NJbbcxvI=d8ifnjQaAJ zY1jI2LL-EZ83M;Kc}Xx6#{l;JKo~IZ+6fv`+v;ei9hSq*RK&H&h}^&6ID0FoQ}wlD-B1 z;t>6Fe;y+<;dOg+yzbl*Hh{Z(#(r(}JO5($A^e}?7xrK{+d21Wr#B3T`+Ia7Qm^p6 zdCjN8e-$}=zJ9;|6py{R*#8oFyEl&&gpm%sGjs^+{CSd$LuB zn7mE?-Rsfz`OZ%Kl(hR*^~fIlBVsecF>bwv+Zck^<0aSqv|!C6`}<|1s=8INra<&~ zJPNGWW>f3GxKUC0e={2YDfj;l+Xxq^Ef)e|tkhYKJU)}UY(LVz$4P)ywD3~gP9t=l_rK%g8Ks2NhH%eQoVZmIC`L701QwWargjL`qUgLrU;_& z8{1|RKad(blt|HEmH5{hl$`yC-T^44Vvp1I7SrK~VOH^~#Cx@*Bc5!EA-H6wYvto9Objm#PKDEi9!sg+ zub%Z6FB~jwD!Kzh zhR8K`t~DcXi;_>5it(6|ew$~(1(|AV$|dgT{*dk*g$ra-m{C5fR)b(t#-OPY&nFb% z*<~Dz7af+WYbMZHkbd41Bv`gHu8H<@H5( zI-pXS#Q&x%>TPbh%6mNs{}=}E+gp90)dcm^>93)am?-@7&O_MX?uK?yKNQ&IukX|$ zHJdf(rdXVq7J2J^P$xmpLoF<_$Cf{Ms3W-0+t>~9_~(w%<&J-~5Ron9%DS`7h53mWB>N%DZC$RvljAU|_3rpSuUA~gK11!tNr|29XkIn-My+#g>_K3!+j;j82ta3QtW;6V#+zal-k@0a^y?<?Jet^(IF`;?G8t}IvK#gKiJFXS;saQ49Vd&7yt|{Vw{7SsBp)yyVFp&^_hR_)=b2F9w_J^xsOEs z0aiIuWy8{ARR5-_c`vap;P9|pkU({ve&(Z5X^~S|{(8iUN8Wd82risUenflKySnif zS#qMkuCRM9ce&l_B5gZJS-A}~PSmz77&V7FGB|6m?!6`&y%_VA+gv}W%gFuo`E1QU z>+|DSzRBRTr$H;?3PM$*ZUcX#IHUveb}-t|p|Y^yH?Xl&5{9*o9Bd8te!DNX;P0o~ zxF>3R*=WD5XyA^cLof_{I~qpOsy{Tl7%k+mp)DdRGdun3o4431=;wG+d^1Vd!5gQi z-BA0VLH*yvNe$lbsk|@2F&}4ujS$}+LR+iQzmL20q zkgB^=GvuhGiF*eqOZbCMv(nS2Hyz-?*o#tBmy={23!7~n@%h0fm7!`;UiFWQp_(S! zw0CDn2YAZ@rQP=hUjf%cr&73D30Wwlo;DIAY7ZQR-2-%!l7Cq@8_TvTFR2i)eD4jC z37g>2&(uL`$Mcnfj58@M3Q9Fa$}{bh$xndJSqW~X`l`-r`|mk#2!OGrborg~4Oi$cHT02%NCfu}BpDDXTnj+yJLPca)7F!mDREP&~!~d@ll+~T?pS))G8~MlZJ66JC>gyfZfAS&+?7r?(eUj;Hb)Lkf+@E0sB z{Hd?!Z4_PQ{U$Xm-v9YGZJA*vWgos_bo^p?!y?LxUBL|ErYB(>*Z6T7whSs|J&Mq= z5%@m(Exth`7s}eFcdJlE_UiHe)!tt5NDHzrYVaD-{P<)rS;`4}ckdA$_0Z0#TfwyX zMWC>-w4Mn@Xv2S@tu^PC9ye{prtvST*ZK6q;rN2y_gt_dJ*|L%~`f+5Z*Haf)H*i;BJKN#=yY&wJ2 zZWe)2K>Y3QMuFgR0C+O_56u4eEH{)hoPg!?$>G52mM8ndk=8pHCinZ#!IhY%-eLZj znKQX~IM&Zb$=?FHwLhTyRD=QynE!hD^dOD zZOcZim_BJ}YO2;cya#3N$S1h_ie5a`eIXLAC#ywTbr#k?Mp78H3Z%UFid96*_FnB4 zw?NiK^)~8&^K{aXG9CPR3)j7LTQ3he5ftn$42|zbl3R^0^{Dn4(`)|n`@IVZw5xvk z(Bt8N(!0{H$DMWfil0-!k~eWUCawYmbDw34CO8d1Cg@ft#I&Ht=Sn>3X_>5{#4x!* z4AyTi1L*{MW6u`qpGM7}7vQ|o^k5y*$LMKsNsa;6>A83M=2S*YgLQWSrfHT!NtvNx zw4HlZ;Wogs^r?LJJ$C!|M(FjVqB?y!9+l+_q13N-0d_7z)+QZT+$$NsmCFii<#-W@ z!hWV|Bh5rmj_g(>zV~c9j?q=C+r*vMb1OkNBG1~WSyq^cB(sx(><_-0t@fW}18ryz z5k(q_7ULD|{dAq7Dt6vV3O!@`CA@bG5MiRe^P!_vTY}0{4~v{~Iwbtu(C=<8bjPNv zD%*fLAVl)Wf!0Fk%~;qh)yvxPMzmPPG4}ais5nGLYJ$sAYyu^;IW6xx$B|WKz$G6F zf}grszxck#t+I8f{o0suFakVVcsSO2BU_LEj=#Wl5$8eRjlu`oraeeU+=JIjxk2>k zd+8*)5$fx@6E?D+^DEuCk6#VjMI$kv8CyS*1Z zwFOmUXM|?A)3*D(4hy7fmY*|s&s(X$uRkVLvu;Srb27NND7uAn(?Br)MX0ktiGJ10 z{D(EHKx(b0+=RaR;xuEGCYPAkinF=f2H@NI<5!H%blBLu%*#GHf$5uAR*u|#SfnN{ zefdqIN*(N8_vlLl^gYkP2*RSLYNW*%V>nWV&gSs=VBm=T!+ zY0K>;QEQ`^nmO;yBJtt1{{s6_o57N-A-2t;OUk%`&+KTwWO|*XWHbVpjg4{TNch&X zgcJH2kHqFR+-lFVyLdE!HBbJEpBGaN_2#>4!Sm^QvhvGs?WGsi47FTk8vqT5@0eD> z0%lH$j~bjBY&;wKB8Y-B@AVXEF4sHCX4{c{maVMVS~X_M(L8nLI|599aZ5h20bAaJ ziG9k#41{r#)$pK!`K(4J$cq{)cd zeG(l^0b?v?lGxi-jFeClI>`}*?r2pG!%Nv+dIn|MRGwZu_GhRpi8c8Y5!q$e-6(js z57zDqviMcH1qys@MMo>eX4|13zkS;`F>z6ENrpw_y4NS~YY7NkiS+jdTGo?WaiU

IIHgTvlxIj-^ZEvmlvj$v`#J+1U ze!p?umiXp*oP|QbWAm|46OQ`92s7VoVqV~$hlzzRF6wMoWFX4135H2t-N zET7Fhsh=8+-sSXcJU+$Cq#mm7`7R{xwf*b^yZamh`342&)lysa&l>O9neJ6T;6}$J z)(CpPw)?6saGpIq*gpg#EqYyb_z4>wX5Maix!KOR3qJW(w!DE~n81GE|BYj~LNJ-7 z4fCPh@Dy$Qz6DGK@ZhKkoaYVZ2#-ArSd8f@%U@y|DJl zZx_8?o@vN$QpSSATkK7MnJ0PQ_nKzJwX+Vv;zqy<(t5L*oH=EYLti}^Hv=hRn`fJJHr;#ZGr6(O2hBY# z8G&|}Q%FSC>Zn^Red!X*u^+g%HGJUv8y%s3u4YSL&j4k))NHydi?pjLZLFG#^rS6( z4~fKq>+pjc*`y^O=2Ozo)*s{khNy;?kCen4#s5wS4f@F@#93kp!cQ5Z(6EPAUuV*2 zEdhsO0XJIl4TkK!r{wMPL zCv!+AxD(bZC`P9fEe_Jel-*{My*M^2K6n4(MMY`=pOjhs($Ev< zXyE)vli?Z=Xd242Z+*w48UC3%*zguDMh#vDgI__9E zR#N$1cCx=zUljX0l>6OAD;z*J8Xt!=h!Zy2;#nBlH@_}vK0E(;?=?D(zadV@*5UnM zK>rz2w$sp1)PV=ZXP;Bq=t_86*RuUq$V7<6%X*u$DsW0icOkDxEK)U~En8kp=G}2& zHId&&NE@b6@`?o^%SN@XC>Nc1#fEM)_OL;~3ZNP5l8z-eQi9pd&=4#qm(h=X1ADWf z=23uYRVMep>tcQ@SQ!Lkl_XZ@JpN|K|51-TlAT?EkL$_QHW*miS_IX<{UAJcIj(kR zNKDwNtUz0}o7(vKQBECXu`MBRnCv9i%azN!`-Nn5B!E$KSGV#2FQrUMg=j5m)Psn` zo-5|v^!naj-Sh;%_F7lx`kYCm(jzJR5@N9w&w+oFG!Pn+98mg!_%@Qc7k*?{XvwNz z!LDFV09b+%K#-XzO%rh8;naM+NT>o3c?O#VOZK^+Drz<9`T1-5@*@@QtDI(LsOB#e zZwZt|##Ix0?tJB!MpS{h;@fNN!z=3V6YHC*X!c6GP6`UZDdx}IRfz2@vHAqHWO`??%a$wwDRhB_(D^)P3sZ@yP-8QViFQ46qk2NZ1 zWm=$v6)1gkXG`kAP*fxo>;`hDOLeexlL3j0vDe&5g0)Z`zJ(8fW33+k0GFuF!_g_( zlaZ80dV_>y)aOPZr9f+4V!l?fvmc*Q`u|ESCnq;JOKy+OAhB@#0jeY?f35WKx47sI zMUg>ae}g^Chw$}nta@dRE`(AMRMQ<=&%tj$?$}~S()b_mCmDs@*`bqEzY3=5)P31} zYML8JZBJ!fJj-Q=5!IIcksW+@@@sKCFmU}fcia$Rj@$Zi-5<9{mtn+TWKxmx?88t>br>| zNc(cZ7w0RmVjIJ=+e#H+yiK=$y9guQOQ~h4T>w|`^rX4y-P0l{#kG;qf#iJrokp=P zt%eDEb`_;o@;?aYUj*j+72S*@u&vCXI39p+f1xjEL!%lYN#5}QKu5t6lILcbFGt0f zjJfTSRjyXi)2bvP)n;eM0_rXnUHTLzz$l4iKr_=u0|ynM*iNSg4P)J|W8!Y3{-!Ec zK88`E-rNkL@+xg`bo4T&E`UpdXmg$Hb8<|c2a`dU36}idYW3#kgefri%iy_W3-j^v z3>GIQi2Snv%m@`%)K2=%A+|NPj;4d#k-|f^2z{%ft^+h5M*;;dW zO5|V#*`Z=A`cq-a=v&NV5Qq#^utd)P7>J8Lm8@Jlr~d#e%+RN4c`*oX{AWi7)e z+5|O4)z3ylSQ8PrU`|JaL@2*8F!^aK1aimQWQ!Boiqalm!+v)F@svsdk$Ir{oez{x zi!maF{@5rbx*&wdZX=&}>+h9EIF+qs#H z4%b^$3dh)7(HQ1^NfT4T-QlA;=xnj**;3zPp7&hZt!z}Sh_9fOkfy!d6AVp_SQNiv zVYDrW0}8%kCN+hpK2u|E{k)z?q<=)*_#Ewp+39haBPe(x$^-sIlAU~}Hg!%Jg+j&HZU$o(D?Yrn9PLdXL@L(n?`Ju$?*xObH6h(C~H)V$PY z&VJ-+Gc(lqdZkxA9fok`G2ABKvgvfP430T(VrV`n-k{c^%5lnvcy?)3`|2DJg=M$z zVZQyYv_A8O-sEtV3n%u@knOnSPJJHPfW;M@PyyT7{lAAtk0H6vlFz8GLU1-|#($FH z`0yQU78MbMn1jF9wg1+Q)LC&kOyDjnXqi7E>B5aeKrK_XiTb5D@tH>B8)v15K?^O? zA%@@I->fCW$X#4}Y8!Z}PunLhdwa9(Cs0eiyOJ3e_vA<7(Y>;yC%ERMv7l1E&)T1}j zV9?T??|WI87`Y(lKcDHtj=wZD8AY!JG>d$iV9Z>B2n*!Iqbr`F zLundm)CHJ0FTJg;RW3LZV|i)4+f)g_g=OMM2zpZ{9?`*;tw@OY!LF`!za&*6iO$ZH z00)PN6=l&&0{4^Z{5Z|v1*Ez{9h?OT z?S)V$+}dvUg!oKQhn#sC&sKGlWrIO6X( zP2j}ZQrgub#hrEu9O!BF=v6hXmhAg%pECV&zveM^=CM!ed)Is<19_HxYdwojJqWgM7MYWr z8E`-S-2GO(R;BqzQgH3DvAxAg=D_+5pFya_Ls0}jSlKhqa#N)9a zz0bU0#z3#7{d4D?L3!-TfJnz(l{Z_=!u_AVNEf%G3K)STE&9tY$qdsw{<>}SnDpF- zo~w|NS1;cKp1qd08*h96yZl(%K5l`P+Z>q<63lru1GPf?*8$$#lOhH0wJL_#gZKV; za0-#<*WvIc9)Bc#k2Hu1Q!7E~ZLG_YeS5LJ-izHVnzw$nLQ5h{YfbVh98A67_2agT z)uKXsCim$hg53hD;A6&kvj_SQzf0SHS=s-Y{{PKH&BLPELH8@$IQ@#^2Pb9O2=wCN zr6ba`<8!miQkIqs^^zy-8#Qb5mO2fWRD#rYLwdu&LM|LW#>ufa5xi7rogIypR}@SsB=J5Z+LZ-Jss(8i+)v*Ma z?~)(ePF5F1e3^Uy$&xEt8OrvSV0Gp0lkEDAtRp-9Vyvj9H^G?t<%dPWW}>31IYv9|U5JV_b?^E~z?dcWK>3YzA}m zuwQ*p?o)A4JW6t~=Di~5p5b?&{S^`}upcLyFNJ$!)iL}P>PGgHCcI&GZJ(6)t@T~! zXY9it{iDGMjTWW4$lSKlV{yjW!IfdQH>9Gu2Nz{n={H51xyWQha`~aTw@tHcRzgFG zneo1u`+eLAzY9YnGvhpQ*b|^)zqbYVKV8GiserLrB@*-mU6i z(QKoalWII?!w9@9l6+&710h_yIq?%&*9N%la+w^2tL0B{tGx3|rR>D7ojKktktCN_ zYB+lLQQlr_;jcGes#`8^s_6pjC~4!+Oj+A{IdvU!WWQeL^0dgv)@Cj&a0g7j3EYsw z-Cm);O*Eqxy6iiOoZQ+Gx83=zT|F+G>v!M1L3660+?3(l?XmMHEeBt);w#iHwU2Y_ zHkhQYyB!fWn{N5|UhK>jibYQ+Ph3to&T7l~J9 z`%#UGd=iZMY>j)9Y1DanK+t3%Qrd(=Npy_%snp{5cLl)&ee(8b$-=;yIB)k@`Y_KLu46eB%esfZkP+{x1$H9DcqT1=tj zsFrv_g(38~vhs|=+S;{4h1BxXg2W{1u;dSW@$WHsu|m}!ZLA1loq@FQomF!}192Hg zMhcLe%JWZmCCj^W9t(EKO{pa<&ax;IUirFLhiqOmIPCSVve9)xWh3%dYc0;&XWGBY zt4>5j^r|Bg@#|e3U@GjSl^#RehT@Q4slPM5?u21K&>4ap7`sO_P11hz;)*Yt04!ig+*yjmu5h-t$DEES%@1eD zNvOi?m%>LEP&$`EiM|@7yKXS8P+fNQXHB^L^Fu~^8nYE9)RTA$<;~D>@+=+idOgSa4v_e^#1@SbS_jP8!~-<*8+3Bp&pq1q!tXGFdkeTiYc*$K zgp6tD)Bq2XkBS)pkbl_8Ku61gmBLjez^X@m+{*9Os-F_;Q38@B72zST*J=()HbDXq z2zq)CToyJHgn7P=X_BJBf6nw>lO#-p>I>^L2(3^blzAAGUpJaVv@F(3k*MYWk*(`;6(#uG9|LzuGi_s{WB3{+EX9zuBr+OlDB4 zzz-#4XW7k^Qaotcz$DxE&kNX9G;6BzSPSLx>=PNYt)QO64_siAr;au9)C;-1CgC5$ zVP&pNoMqffccqGZh_NzsvPQ#>g zF$<$ph)%ojB85^D+L)x3-k+cL*=K`$(>_=UrH!C$UX3RA_-BdJrHnZ4Bo8yR2!ZSL z3&T{#3oVw)=|S9BlI3(O!!iL5@VBl&kI)q8qz+e7!|@KK+AT-IItNCYiT8N1P*S(V zX8f<9rtW;^Ja-W;EPb!$)D;25}W);1|5fP9>?ZqxqJ6rY1Y9(f^rrD@q z?|W1-l-tqBPF2~YB;h8l2=LmH;7hfrRKfTW*k*diFt*HNy|t$Y`Gx`nEl&HfFvSWv z*JUGQSnreTX~u7kGI7*FP~`Pzg($f=@@ysCgJ(8Z{Ux?`;OkYqv5q=^rwVGFU^y}q zs~>RuS2j=8Y*b<>bazLIxi&Vjip-BM30^ru+3!4=A2z~d-Zt;}CWS&$63WC|;q%{( z?F@tGzeose*R9Ssi`D+`+vMvR6XNHyjyjU#8mJrvgov?|4*`M*uVlNy=!pgA+ZIiD%l#-TXwO1$&b|6|mK3UQ570kh z4HkSkee@Jqxp->&OuN zt5YhpEAMB||HTVq;Q;6uC0sFU}D7cuc7P8Mw~#z8x;EiP>&BA+H#nSFIkU zP7?6jhKn?};&p=Llzs+m6wRlo-^4^yqb1sue<-c`^f88<=UbO?0LmLnB(#)#!D#qu zHsd}tk*S5+cS8RA45R06=4rJ>pOL23PfR8A(LQUL>QS3DpAV+C%2>v!eqBm;U<6#k zT5X*I=CnX+8F*QWDbh6Pe!I{EG@aNjh?tppkv(irh?->A)I`dj+csvB^KvnG)QhV`>yuWo_49KDRh$8&}H)YnN2E>$P<* z$khVdQOY&*!xT(J-gvwG!0x8y{p_l~Ge7Ev8SM<2C*LApb3{t!mDByCsOx33bq%I& z@S8|)YYMh6*q*Cxh~YZGRQ+<@MKs>r@oy68-ypklMz*uWCwdvzL-ZVV zzSn2|X9Iiw7X`)YwtF=mzg-d6v*ofiYl+#}S?NWL+Tj@`EOSFVsJdKTWhIMIcqb~U~v7J^2)K3>(-h}7X}zAig!Q4w$fV7^wwr9 zyFqpL^dn7whpHGVk&citsXE2PY7ONwXD4oAjGUr|Kp+kWz&oqEBO;emH`)N4HobSl z%kDqqSZ0@AoTGOVA}vQNHhlAeRA9FYFl__vA}K%wtH=ZO;tZ*n+?jQF_iKtP{E>)` z$Xh_;I;?+j59q<`B;VxI{+^I+KR+20Rb&gJT^q&x2l!lEi3Z(m=t8Ei=k^Tk`!J<+ z^L@LF0i8sls*{RO1)0g23o4#@APkBY`^v)^DzCJHgCedxA)LJ!{^#|rz#NAv@Q*hE z&&PgjY&SMBe^#Vbw15dN<$c0@c+01Fo|ZX5zVFB!TUyNFYQn_pCrF2xaYA*IisKi$ z`qvkr$NQ{z6QGYgH+>;S>aeOWKbzdPInLQ}gS>5)T$ZZn zb18CsD{B6SOnzyKITlsKQoV0#ylc{=#T*~9N;{=7_RaoG$*c%+F>{cdv?j{a54p$k zKA#tRTJyg}f}vc-{?5E+W@81D?K7^o$dSCqK-sz@LvP}9@%$@9!aeWY%iKfvkLB(bf{|+9-H`G zz9iOK(rBIxrJlXC8O`fvyR1_b8d$vkk*v9E#<}!$$&)a6+$?qU2iPK#yRU{?RA?qkr37E?wd6(W%?aCr6mbPR zEkvBL8P$zj;Wsm33FH9cbrErV*?AZxcwX_%RUUO%>GuKyAmfAZR zBd3KCNvOZ#1xGO8HYw%Pys~>mQ+#XUoWpOoIDN^GBc~^=LfnmT%&2lpb>R#70puq88#iI#t>VnUXM!-Uj}tr)~cX zx4KnsdpP;`nvZ6P14~lVZE_*K2Vzr+_qtJIZXu3Uqz|O|u(IkMuN-rvmf>n;=1!Ig zClirEg{2GW=R69*D()YYiJP~{f~bLzYc`(GiF|MGIeg6&gxp6NXo7Yc4MR8|V4Pr- zqtRQSI&`!ar=XLStrx0iCRd}>-9kYomWW>*;k9`5+6&xL<9Z)GRVaVg0WU-95lOj- z_oe<)$$sSaT%48jsJyP^r?@MDxnR!7N1#ujPw=s=67=SQbT ztDE{*(M_4yesUEWCnF&RW4{F|X*BNXF><#GY_23ZKDEgT070#r&{jnCnxFtWA;4)2 z8^`{f{1}vU4#TD&v0T}sK*|gX%V=xSx05dI6A>lD9(!8`kK`z<_)rT(oZ*tQ`{cvj zml_;MoGWHlzZHIti}&Qu4cvpEEzUg;wa&}4Ll!Sj!K>H1qvJ$Ph` z+8~L+%7!~EF21l1NNT;!K!wMSo30xz%ISiF)1a3K%7W(NVVC7OU)&ca>~Qp5W7VPUF&D>4hHdpi zl-17qQs`9j`(9ObtR^wEH|V%Vilady|LKm*AKvm}(X>G6Xn#X*y+K(CcXx8x_3%KO zhHia(dY&#>w1STvF8Fbeb143;)Jf8T{qYBIGD%+}A7cgc@*^TUbfUIatuR9u|7DE# zb?ye?{N7!a=wG>kZ}g5)$fYqgEksPNp){m}i{T%(gvGJZzMLso-GUQL5~w+8X3LQ&wCa&0uQEX9D^P zo(s!@o@RLR(`Fr>@CuE$XVJN#diDF_vhapx3!R$B@$ltXRc0JTih#}R$qPr#i?}4| zJYZNd+nP%C=_ZN(xyA32n*Bvu0j<1?1$cR!;>NC>GLmXC;TS(7u+zl63xq!b1XCQi zTbB2^$7{4E%T`YPWv#ti*x<`oCYMLNRaGJi_VGxlBobW7pmb1L2Vf1$CC@H9`$5f+ z`KR3-*wj1F(PWnuh)o4fOb^pctV~O(%yLU$@jFY^6~jg2cZL7cSpO^c)Ax|=l$`Sfx#UMz8?h6?>49L1M%fX zmd`Y%z1LKG4#m5SBG&g>%=@Ld{oMU^qS)ncI?iFf#tS6A<@FT5VK7ck@LOqfkl<_h z+hZtM?1O7NbK2g?V0~==Go@tt+vCRD18aMf($Nk&rzIrVU0~Z}l=q((Zup-kpR)mL ztQj2Lta(IZQ8^jGUU{hX3BFV9JapVT?_K#!E{TAo?}bIY{BrM0CK*Imnq@N!@9N@4 zlPuH4)sH2Xcyi8Wb>B(K^TXs`^<7ZaC3&fhK(rD`-&i3QF!!}{{=7X5;6`dd$D-dH z9^L`au`Y~FTOdX(Psy3=)^9NOrk5tnNMK*d@9km!)x$2Q@NhBBQS$4>XIFFt%bj=j z?a+>p_9>1^rEPY#gn`+Un=cJDI@-0#>gUb@V146uFRwWR1{OG0Uf5Mk8>kS9oGdk% zjE}z(kC-?sn*cI9&6_5(QqHVHrfBoJ5$eDuDz&_rsxKnen%qkiS$5!&WrR}u=@NiU zmD>!DFKG-u>}KqEZM7ASj$xKs_8ZSHa43)x5X5jU*jw^Z-wG2KgP7cq)Zh+uSouh8 z`)q<>;^+p`+W>#zX`TKshe#FlW7F2y+!s0XEg(tO_Na^fdXx!K^ThMLhEFx@Z5vE3 zt(N}M4jjiSY6Zc}ry;?LK#h^MWoM>Ae)Vdy*?|$xE7*932@VpbGKQewEh@C zXsQCWR6k}_s2&->u=rO}AyeGKLKV7%0eVkIhkFCr$ari_+L4GmZCiV?7?*!g#h}Wu9d`ux2{E| zp=~$Xbut^$=JA!9<1k;qpe?%im*{Y%CYK~cyo*HtBxYE75?&Vfnos=YxaZTHDzg2 zCfmw_pDz0vq2mDw(&0ro)36tStR_LR`xN5z%eA)$vNRnFJvWrIQc(GPJjw6IKkm@73~<*Hvid5De9Px54Uoa%@0LMv4zS?M?29tb zXKs|lo(ZM5sYpJraRYe`OK&_^hXiZwIM2H{_$d3f!74#XD-=`#{rp7&RrAq~_417S zx~<6q&Fc(-cvkKRTP9y}!Q}bN^LX>w6pz_Z7)v(lSYBbMQXn{*I`pFXu48Nn&K%Mc z?-8P9iKWWjS4?9v(obTSCoT(xlQiSM;PFpJu?;PzsJD^iw-jzWSGA|HuvW!Dqzm%X zQ*5ya+<$GRBlJzDYuXtDI$}*IB?r-9oCv=dvXpMSVN{}?m0^{Fewn#>0!GwMA!j}@ z0;V!r-**1Sg99#G<#T|=AGS=~$R|0*Tbjr!=+C}=F~VDCEn?|x7P)kZcdGt?PXd+D z5KU=Q9{Os&Z<$ZeJ0c$16|x7|M)}yiV>cwm+5J{Ud`|J)Q47;@gJaJlMfKK&584Y+nFZOX zt)YEm6{rYq0Uv#W+t;k${+bw!hOFN=3@Rq@E&RFQoto#*t7B@{ zy<2rOgH=qXE822;T$ z?`OhcFy*YQ#wmf-a8J*Q>vWN;*ZrwPJUwSgOY)9++B~_Dj`&H85{?P126q|9;)Ip} z{fq>`kB?<=n2`)7svde0w*tRoz?{RMEE6gG>Q0~cDj3^g zH@QPKT)WF&>dbE3dral$X9%ps6)S@RPE<~&st3$BLos%YfIy^F=d zlfZ@#X~B&{5tQ&fx z-+6&=`-N$Mu)l+Ns%=N;83_0EM1hG*Ocegmz#UjDw;B4{aBGFY1#frn%JCW#IKsx% z^^TbXO)Q~VRky()(9k*xbw(&5?d5r%gD>RSU;OLV@*Px)f3spPH*G78Q5e$Ww;mGfiUz1D#glA~ z&%dJ%gi;L_9v2$pjEa&WvGDW3I5hqgsEVdJX&#qOtZ4$Gfd>kJC%Utx!TDZ+>E<+aF|=nkR$CMyQ;U!V0?CgPcxkPiB8{_XICsWqG}kwC|i=q%1C@a zc}-BCK3G9KvZlJPC?_ommC`hs&P+^(2E=a6Ws}WC2p6lV4x{uF`xd45N=WSBonbvBlIF8_EIY#N!!z>GHZ_B z#woW;L?>b}0}9k$)^9;6wQG)M6dvT(+CTpp5`ymIm%zj1ClFUlb-;0YayWAzNCRSf zILaRTl(q%^!30gP7IR&&U8viAvg(8E&8V`NU?kzYH|HxYM7?RLjDqYVy+Ly7VF_Nq z$R=5CrUY-8ABw*#O&FW!KsDfetDIH2M9}?BG`wR@0;*2}WzoyQT%AGxOvyVumhXY{ z-uN@kizUmDIXW=BxF$o2DASjN7JWzOVj%g@{_e%dX`&16C*KMxZk{s0TnO1V-+hEY7a z4ZKS>RVlURjs?;VPwgBS_{&1{vG7x^Eqzj#jqKOt(OI^BM0jK{$rM19k=r%FAj9Z89qW}|}pL0gAVC)RXh9v{-U^M3W zR`IFfcYr>6lxblgn>FPbdWKiMaLx&vUw;GuNvXj4!oZV5f1wLL|y_y5mtN zIi1~?wd&CJp9(`7QM4uG;! zmAhJ3ZCp<^X^Q`duKy~JYnOIO9f8PGVo(MuDit>HwkZZDu=sw(k_~Hk;^%xJ*x(T4 zt#Bp(-huh$MnxmTx$Qmjtq1>kg^|qnaX5FQDibYQ03wq;vvQRc>q#MyN7Q<=`_`#_ zxw2ruzi@+3s*Bm{@a%>d-D?q$(Y%HmZ*QWs`G9ILpwXP?xZMcNrRR-nU5Hn zM$L3}(SuCisCiUI9QhRldZc$1Q&fSEQ}$HMxvsqon-Rj<&E z^7J9ArHIe#9GukPC&s`k*ZxjGpsXLeVYE$I^dQeI?z%p0Km?;t59%^Mf6;3UtuBqcIu%e)?L?rdd89+kTdZGI#y{9~T2Ml8wZ z{ANp;%zhPly%s^_fJ$mP1Q!m`4A%7(kFV7dtlVOFx>*vq6O^OuQDrdZAS&E zif}q3l;4eO5hHDt6#SVwzZF!HhvER@(J~P>oWQ3h9E+=NL2VcyF>YsJGaShbfptcO z$$ZD|*@~K8Y~_9vvF1n^@gEv#Y(T!T5cj4oea3`VNStupn@}=KvO<6dy7DFR`b9+L zZ_TO-$=ZJn(LWmcf*znfK;GtXc{jj--IQ@K3R-)&&Kbgh${pfu!y z``P=0Jssh4IDcQb!Rk47EGF2!qH09&xq%6f`gIlyXaDBxo;BP1phXINN7vgw2PM+( zcOT6J9Ds4>+m>VqnR|KR^uLI@%XI(OQS1L2nxja3Z%9$GvE}&PbDFp5xKIkOa0!`Y z%wA!m*a70o^=-$KUxx-GC3QKgHS;;hhH$YRhopid z8aEP)Gt@o$r`NfSx!>IczVn39F1SpkDvZhT#TGd6@;&#(7ufC`aE6R;J6wb(Ejad7Ph%!#$!0$Ici|Lf+}-i*NggVR^Lc!SHVLBfG_= zAJGelNTk_!4vbUObN7&Q6nVi%hyM|#Jhr^!dQGqejHZby-q3H}xn<^o((~S9RLKpZOK|@ohIMpkA}Y*t-xC%gw)7 z6R-N@{d23$F@vVlsG5Bb%C7gmvm@kk*wiAjTK{J#2b$PJ=jZ1i9vzh%r1>TlZ0Dg0 zc{m8_)dxp#(U*BJ4DVd5E3?5IXrP-;Q|w>*6*z;C^zMb)7$?19!dK)=FNC{i|ux^SiQSuX>sQ~TUn=8L!Dl`Zr1eRz{45NwzVHUEMS54t%eWUY~*BL#WIsU)rhiP~4 zKw)8$*OlR%yE`WT=W*A;)vr|HqIm_O_YvP^UFVps$L>d?&6i8LJ_`2?ga7U7JGfcB z`J&d!GA1B|BVAaTo{vOwQ~86fTKYY~RLwdF5D!jCPPXgY!YT zk#r^e%6Z|QThL4%wn!TPxwx^T)@<+|bDo-A&n*kEcO+-IS!DFN30}T-sO&jn=@8TU zc2s=L8IJj^V9(>5dcgU)2H`_fqKNsq~?b4sy4q{fMZ4aY!-x_=;MT!o&;o2xXUzEG|!gAlPGAQ??< zUIJt~DuAbzN^{X4z!tV0EW{14sO~Ay@qg0G_IaTh<9u#%_>`|G0dM~|=3eo#Xu}@C z*d4;d{faSUMng+l7JXV}4eg&?6Eg#bZ#zq=VGApdtCjt?sw>E2UR(C^R*V8v}@jurGbBjIFE>`~`% z-@O6j+Phslf$58tvD2vkZly=W`q4Q9k_*_Z*{LQ={Jry4I3T9gL^XP8w|Q%@trpKXA8I<7+t1Yva5mEpEr07f7y%k(|369 z-3G+g&c=270t;;Y{`_D0_Dn)XCm++*b}A#8TH$g;LgRAUzaII298m`v@8N3bI))N^ zeQCYlK?a|8$X0^O|6BC6>2QcBD4-Aw`Lf(v0kq2Rf(nCZqjPe9RfjQ>*5;OJE^A;+ zOK;O!jWnkt>G!0n`*W>F*M2wJE9LC3<@+Yp1w1J+Ad48cbMWxWI585nl=%>3IJ-Q3 zgEb9`iB;YUwpfC@<<%Zj*&L`Zl#)|0E0eKK1|Nw7lf?#aHlJ5AT|&bTD1++v_XhN? zt>N~QfKt|B$H^sW7aHXu0IlZ~O1K;BWGYZG{4ePu>Ma==U~LF&@aXoyk@~E~)~ckA za=HrIrsPGZIF;-o4W@mZ6V+y`gtcPSfYzW-p!nd74?gzlZhp-wH6P7JO)7M;_#u=a z)F(Qe_;+<&fF;-)Q(r88SFiX}%5)U>$?dY<8C^0;QOCj~O3 z*Q{>wXIpV1tXYdeyypQdJ1aWcb@b#eJH_Qo!r76PR`WS~xhd_YK>gi?O&yMzt>4K$ zQ2do@kK4YvjNb2_b=5KjdDFNh%l&Ok&Y)5r*1XM3$Ro)na;9zzHs?4DEte7tc@!1N5m4&Uo2H zOEi8a<7eyQ9;cIks}l>SNU$hBVxB;ND{sW+osS>sfW?UcJ|B&h5; z7|gezFAYQRb(K>T`OFXqbSk2(bqum zH8=cGpYs;_DE4|+iCI$rZK zjz{{8k^2<#aJE`l>vv!21=88J2H!hkv%KZ`?;O!dPHm_uxpkA;vM{~b?KVz(02hwG z)=jVzM!m!)>VIzEX84nC77|QRzn3;qzbM%3t;!_H+4{sr>_%tyG)H}M`Y;7zT(eT9 zfqX!YwDjS7bzshKWO0+Vgr;2#P10y9*>k<6iqVZ`^%{-`q_vWJc6VJuYR2A{G@JVN9^7Vl;J1T`J zG_1sX&wBH9!*2JzWBOhjPOEGlv2-!g4CFs@_2#@Z`=#dkN~JvYFNhaAzC7RONH*Ip zKiuOEUAA!n1cu(y9E0B1GOK{fN-r*JE*?p;{<*4@5P6+-CfBM|lxJ^bgR2D>)zLOz z?JZBq4wBW{-3A4vpt=He@Othz?3VC0hEdTnd^;&?hntGkZCQNhh`O4f21<<%U1IC< zv$YPQD>=KR%G7z!_p3T|`Sz0aOtqp0t2GA=>%I~o-8z;*M(Q?bt|qihJ$L2jL8`&3 zWH5ezHuH&)8qk=kFLS5nKTg`e%yymhgP{Jp`NfPBc5{&(SSPk+8KJs%Xo9os|2wpKxU0d%yS+?*A12 z@`$7RD{zYMvXkQR&;OeVG;@2qeP7wRpIVLQ^)<(=T=FYwBul5{&le^lk@ste#i%W= zy5~}&G@7Wu7mAOXH@=v_{7yS}LBtyZrOP=TQm^i4*s5hT*$Oibpe3o&nu633v9bk7 zCzD!vlH;Wr6o+M=qwI1%`*t}u+)ZO1Sy)vl=IFSoi)0*V&RmxM5?CWz4Y1)KMN1pd zmYE8Sv|({=k+SRXzwCN*wMDh%j_Z-nNYpZT^V>o_tG8Q!>`~0a_hC}+N#+u2L!pd%6-@lY*6;HMk3fkcO>UKkUwc806-v z-+At;Garxp^q>T!J(m91cpua2C)?QLl3J;tr{2E|Bah@lP28+;2pSTg%||h>h^0TG zzYm*HPEz%LAjVmdwpQuVRRTE*1S<-SWuE*5#5L7ZYm9LZf}n=FpnATZyuW&s)uP)- z87k!UOWWULO8O=K+6*K8ARneJVW!qcR4M6ND;e?!)fnKNA$v`TrSK4l25fRG>q}Uy zjMtZF$4#Dp4Uy@+X*%h4o<21ZDa2HSptF{&^pE3+&tFD!S8=nHGw9c;PskSH69HK%9%8#6v-o3}3JFs%Ji44FGPJFk$uawy;f7r+)rvhpJE z{97l2*6oKtTt#3As9yB7si22I3%Cy!vXr=hu)6k@X7c2ncC2?Xd-VW8j;cjnbv8ef z4;M4R_#GWO+M6d;2Hq_J9FeZwpTA=05E<)A`zWo00j#8J2uP2$ zCmEORn(8$k0dg^T2(U%+lS4e8(V1Q=i{;Gh_MhhD;VGvAuOxIqU5D$$|3V!1nhWDN zo&OmvJM)V(AO0Hm?kByhd0h;(ZPe#Pvc^xQrwwYnkxlOhyi4yfI=kPyz_Qo23skd+ zElxA$>i%}@7af)(f=#{l+o+Yd-!$FRWgNa4b`a_F=H!Q0>IP)rGV}+fJ?%ZNB?L4o z1x|N{D%R1S`TRk&!pZKZgVwM#_alDLtXuF)97h?I&sMQ1!fWf$6YTu=J~5&~2ISK) zK_)v?n!S-Dz%_+6=PptWTq_lUX5u11K4@LCKUT}2{Z5fSK6WU@&dX#GQ2YG+J>J&7 zCe080NCn2+BWs!p@=UJtNL(%%>GBz48C)nlEO~8ZhZ{R>_n9N$+K$Ud24&-6uf6J} ze0g=YLjwXj2h&uSx zZ@uzZRvg_M4YORj`BNhmNR=<_8qwvUAr@R}Lf*vdVf<-VRD!=stM5J^!ePCo$DwEL zDZu0vp`cU~VMguiHC4B+^1x@RA2R4YT%!V}*hy<;KQFkgxz}giSPVCR~KE z`p7iwW>YgMhf-xAhUSEw@*$;AI!fF zYO=P*{&emw4>X&+!<|pE?Hfj7ddtYkP;YuXUV|A<4!0t4Y_?=1PD1f?7age$!2 zvMJZja~iMlv`~Ud*QX}F!W5Vddn3T&AR35mf9nL|(Y!ttrMh}02Y7Pv&#=5izmo~s zdVIm3Q;M$0kgytHvibJQo~A5 zWv?;J!LHByt++a_^N#unop82HY|U$V`fhRBukPdaQ#M>TQtTi<>gk!j@WyIf3uh7R zdTp{);G8P*IqS>lM$o=XU@&*~H@|fAQ)Jt)Xs_0DfIw-4+wAfzIorQw;=iR;Z#`Ea zivQC<`$ap*Wz{T=sI9f*vb%%0_nzAShWcf^=S>C=C^vKy&tc5?zrukxL0QA)8a?d} zxAO!Up#NnY*U8@J%kL6~$+ZN))mii`TGrCdy<9`WPdf-p`94%Re(l)^o27oh+CiAy zST8x}ysu8MmXRsK(o~AQ7LQ=`5fZa31jJjn%ZQ*b|O_Z z(XEo5V7<;CsDW*{Aht#6QC;z*S#m6CG$d7=yBrnorh7#t6rv5$Tqn(WHpdvI!G(Dd zUuB(+D#?83XZi^jG`LoU=PzXeS$hOoaRuPI0_~@`qZ4ISxBFCS|^ zr(HAfHw|E_+~1D*dajG{@8vBF-rTyb<36P~$JK?-pGS+u-uJhA3J7{`v8qTB6xu&S z=BElPYSdcZms@gOUJPEsovt1A-}Igy?H+%AfOaS1i(LK&&0JUf3)dyXbzHC#EH!5~L{Z4k1$kHtl;5 z4(G4BF`rSTZyE>1U{xtN#P@_bz?*sbjuD2Eyf>shWuVOFq*hKiQ6d&JZGeT+Teugfj2+4ac-=`;7h0^zyQ=# z^GVF@Zb3uRfMfc(D{(oenm_P7GXcBMD+ik@%$1amIz~UE!s5~}b-0FM2O8vum8~hH zyn}?xZSCdH>uhA4Z*6;^_d|FD51xqMjgOM^l)0{bAmtBKL>VoV7GesA;10M#sc z^WB6$$V5i8JvN0=f7y{`Mc-LQCGD_jZX!VYR!l`0HWo0pngXz3?bjSRX{1QF_BHtj)hxlE~fG;@E&Y zGe5#4EESiZOBN1!s7Ykc=Nwb0Q zV7$EiQGlzF0%=;yRYi{_&qJrz^CYK(q{V1R55uf4S)3f4fA>srW{y54=Kf1ZLXm9& zH4s&*^8%u5T4`{9A(dk@vebOPQPcJ2&^3fx&E=~1H@;`JD^XVw8-&j_9$%J+CDKVt zD5gdF0=y)*uVtiHDeRg#ow1YXke$*rJgz!K9zm58wr4Y2CHtZn)-%V!5XF0#8ErrD zQ3GD_H;e#&oErjmaz-tzgiya^FJU?6sBYu^SH|CKgx?LA6m#os}Vq7-lcQ zIrY^TU=Vw*`@G123q4?IVJ5!|GI%jQzJ=p{1RL2B-%W63tq}9W5JuZlca^p;y7p&~ zJwj9cZe5XSqW*$J3ur#~z3Jn!$48V1#07eF=a&>t)%(KtifYVj$5DII2nOEpDRf~k z>MqcQ^8b*NjV1SOp%sq&Lb{!P%k~NUYfVQqFE&LduKlD;KIt9!H+#yo7Q)<@cl0l; zUhqakg5p+y?S`FHU*h!1&wWibpS_6LvgYkulSOFG(~!k3r^9d$x!eYiMuO{v&*x00 zcnX4)s8f1A0^(KN-hjvB15K!j$3h{q;jEZgc!Q4kV0XLK8>g(XD8C7kGoifcjmJFl zx=%QL(={!Fdyj^j6K!Tt0nuf^-yB?ALrhEGEN z)k_=QjmMscN0;olQwoE(nvaJs0P)yzO@8I;a&7g^EY-)teXpHR1=}2!-?9QPs7Jtm zXfOKbsbP_njXbv)wB{#Nw}tbc|G5nBgr*0Or65CsD|pfH6$w+jMWXm!1-G9!ujjbr zKF9p;y)AYTj{oJjkoV(hFHSe-yKvEx zFwX=WEV1So2BhyY<$&)%$LlWxt~p`y2=P1`b4|!VdJ_cM9awUP-kaG-mLLixHTR+W z3jH|wqLnA^&cRgjXqxg>Y=oU9d2LtnO8U)ypr?>5_Yd_VP3n0y{;^#z=F}AJ0(i^T zpRF3BY6t0VSkh`fy)I}2Q@qz}%23r*?vg4KlO5@Q z5Txw*tq?NdtXx|d=Y@PfZtp7?C?dR8M?oAnJFr#m+B#FjKk?c2OgE9;B{xB$bEVYm9KG1z`wJ!Z#@Jm5Mg1GbZ_uEyyY-@J{nj) zx665re?EnL-21;T?snuXZ`XQF4zhlv|KFv7=>amjy@bhU7l9X4)wmE%bh?@f7oBUz zd+7V$VwR8&gL&oO$A=$%Yad0L=``b6$O+;=v%R8!oWaKdLnI<+l5B^#|4sAPrQFcQ zh2lAWj2E_~{of{aC3|kd0MylhXs5|A@%z(-!KwIMq6DB&g-r=c01Z&yL42%Q299|R zW&syrjqF0UJy{+RPlaMwTRKalGX<9fQYMV6MDnaBDrpH&&A6izEV2OXxNQC7;Y1uI z121m~phn5r!gShG_pVGdp|HLP92Hy3%F0clSb+t|L;Q>PhoH21N~H@JA)H3(?`GVs zymyL5pQuL)(E1CZb5O6}TQ_c^p<}@(Fhwbf(VU_i(#8n10(WyGNi<~=W(h)corXQy z(olk4t(#>!Bb#m0ZAjee{O&Xoua4uvkZwV_Bv`8u#fIP!*GBxd>f3dWzu6|7g044~ zRjHI~Xp7u>g^OeKuai>4+bB_mC7@&eSXaK1WnY;yi6VPlno2o58WISuHz1a#f^nTl zK7X5$nLB@_tX^u5x6lq`ketn?^eJtMX*$Y&JU{ZGwf+XRF*)=D&8>rnPznB_*+J`P zMHU8?||FPH@8s)l2lCxmaOAl;fhuC0jY6Lqv@HKSV@l z$z9>38>Q3yD*^>HfZJHuH`5IjTT5;ap?Ub098qf5mUdlR>8r0i=h0Pw3jvsoJ>;zt z_^`F4)M_U4vJ7*eS{Xlbu<_n6k~NpCt$MeVt&e$~CRaCk1746zm+xc+b#?^WjH{Oh zwU9-eh9CdQl*vx!Rto4zOX0Lpqg#KsMBVqS;!^0m!p-j8j0m@u=VQNg^HG;eHja1( z>GOe#vvCD=f4kj6N4+xU?_<1rjgS(KuG>MQfU4(5_Uq_73P_1_E?P``z@N!)8>RO1 z<8-tI7loz2#daTq&~C9KCC+Jh)?R4MZb$`X+dky&e~LWtYd^C6vea^#_f#y@;Q90a zK797_JG*kWXd>I1@oidrvx6RD^6o(I4Z2lom2dqMx}&TqR?=rq(yGCB z;%o#nmLAHh*3u9Z*h%Cm>gvUE|H=Gl>vX2w0ol7iCAe5fM!M02L%tpHk?aeu9O_Cg zVRtSMgqOA8Yw!~l!>oa+kZ4S=&BMWGs7j!^J=s=%vp76as<8~ya-|@8gvii9Y_hr^ zeb2xhrO3oza;)XS3E;1N_4C6~zSoD^l;Q{#sT)I*1nrIRLaf!h*Dkp?TYg?BJyS4D zz>`Kz2n0!N^qK$UR#cC#u~5jXcy4Q@pS&Y^E$U{lNoc+OK?3IruigpOrfyxM+e2Kx zsz^b#-Z#%AfTsE$_d^gUDk)pM8VRNFwJ}E z|B0X?;r`@0*)s{(iT+J7h}lVaBOUZ;1CAna7cb9${g+y=>3ui%JcUa))V1tXC-*rd zgI~sJKK}&LR-PCV^TDO}@6;&7-g=j@a6Q#q=M=1(?E2#3mIy0LF9y>u{1U4D%;jOh zJ8``a37q2s86io_6oWlL6Td(+v4tWL(vO(S?b3(eqMPHHSIOVXJ;4n%0-HL;X(g1| ze?|sh7-yn+gxa=>Q`;{psqnMa(ug#ehq<3pkofitk$4HWY`F>LSi}v1g?Jy?g_xtA z?dSW!@wogZ-iDN#9OvUr(aRggVfw|2wKu5ek(xldcU_&(N+?XO3qa&**+GeAsK}nF zr#DjetO()tg5F0O;$Y<$OLSA)I;L17g#PU;pA*(_h!RuIqN1D*UruOiLHg63S@tgE z8ucSMiK;iMRRh9S@+MSYb73KAN4 zDwLrYepA54)6P|a*O^~HnMxx~%377~=$?1R1E?I$V^7&43O8s7xvE7-6 zQTOj!aLpP(=w8t6QF0(+=U7#v6;#*R-ODysK25H+R;JCGnn3=ZS;go7rsu~+@+6nz zv|r!zw98d5Px$`MN_;o*NazW7i|}8w^qZzOyu%~_5nXx=H+ub!@wbV7k^P!>pL{cV zIdY!@UzS{~SBLEHYeCSogdMrAez{xAN9{{6#2&(5N`=8BV)LHY0irPo{5>Qa%2?l9PpfvKO!Ghr=)_BWT?zlpnL;7ts1hnnM_fPmKU+ z0`185`ynRKZSD*8z2WzJZP+OAJSYO#@<3O6-vzp{SGHMg6~k(_q*E%2nHLoT?3sia<5ZSbs48~e?fTtKDCN9i^(%IO z(d5N$FzXB5h*uK0DPUzpE?_W#7yFOB1PIJmf$EV+oOtUr)BCoR1NYI{emks+iqv>V z6|05w>C;Y?CDd=wHhf8g>l@!qn0j%F5+gJ%2`1c`#opCcN*)Si{5>3loy0`p8O}r1 zzJ57#H4uf;7H<{m1a~8f*rdci!#c|*bz%#>Jpw^3#}gP|>UjRrTwU+*sW!{T4I(M@ zuFQp*b*bZ})(u+@BdxM;4rNsHWq7TaK~Kev zG|D4j!&*g0v*~2R$Q8<31(8umP$KJ2E>erfC7hVxhhBKpfUyQa1Xypp=`%5*I|!jq zCcS5_&!V``c@#-|l5&!1$uHE^ zmvw)BMVL{Dfc-#My>CfA(cK4*3v~kn0}2+=qPVeiQC3AGcrl0Xk+7Z(Y&GPrj7S%) z{rW;`G6G{Qik?3$u2HVab)Ujo$Qps5-l=TYftx!M6mo(vBR$T%@_ZIoH>8U6KQr5( z1KxM{OnRX>JVlAsT8{&)B->&KsJ3dw8Pt<2k|M~o#I-vzL$%Tk4Og=wxEQRVDCvr{ ze766p66Wa96K~~X%)ykA7~dyR9A>V;d=HbYzRPz(}C>%zl~ z0cM?T^5N!el8u>bK<*zS4kYZW9vwjs((C_a0fb_+*1R0mQEq@S?W`JZHs0($+c{VzQLq&(LOSZ%>-6yluL<-N z&F)eq4&R<$+9;v8^7(CbKpH;wKIlQlg)ZP*&rV4^BoPt|dkNM8Sl7k&m#?L8H^NX^ zZxlCiw2IeCOB-SCo%QWP^6>OV-GnbiQO?z49+@Wg&Wv~sJMg8hONJ_-J%*PFb$-)M zhlQpgr9u~7i3Md@Q|jM}ikjg>3!tYG((DEfkph+D{56Hd{ZS0A(^5Qx?k=$WSrpm; zg`ytsQWi@pc_`cWo3G|5{+|1nqY1t*yIkEK`>$!QQGgyc|A(pZCbuuK*sH4k1hEMQ zuFz%6pico$AdW4#hQ4QUv;Ag8Wo@^%MUH354P-E}UDs{)<-*xnmEYT*#dbX3n)BrP zmwG0VdZsPwRQe-XL9e*2-t!!aa@#&=U*ksSrRvDMQdY*X)ZR=K{H ztIpaB)>NEjdy|eRq!AQc5*`nCAR?cFvP)o*wQ`!M{m)sBFA(K9AA4z74j{Xa%(@VE z5B!+wnBc+|&YA?!W6Wf^8HT~Z8{yO0r4?$5nEEH#ygk?=?JzkU~EgLLU z-30{U|L(TGjFp%f)qYsHY;8$dyr+I=iMC_cCkk>n)}WCZR6|0kA)0oM`7or=K^$%B zkgRnkkMyq>yLb>U4lgl<<4~#oS$i`_Cj8`=y>&Umi+e|UHbd>O0Rw?l<}&9WjmU4R zx8R@SA4I=cCJxUs=;Y>l-(1nuD#`=i3tq{0P0dAj9CJJb;|C^lIXi%I-Y2@54=Z>~ zxjAH2w4;XJa1LqAv6ekHToB0)({%Eb0#9M)aA`SvY^*91px1*$K{Ua}h_h?6`#VhzVHZ^M4{BIt<50!RQAuzCY_G7e6Xo4j9d&J} ze)9gGg5mGU5+8uy+nA=!St) zqDL%Te*VMpcjx@b@rXTK{xY0J!mTKU335qt9xO3^36+eA_j0^1X3I-+UYbN#WN3PS zlQ8qM6pICH+4m=FOZhTlnOOgINDq#vT4O2R6L9Uj^E8yF!}7_ygVO*^nBf*ZClG;? zC8bB>xSDcx8@$*jX%Ve;K&;kDWOCLlEs+&CLSp}?=1RG^p{dkvtAsMMR)9#!O^2?yxo%C*^R*4Z1j-&R(VBhNn( z#auwC$@P_2Mx7F5b>9aY7md8^JDx($A)jS1d7)mt`UBlo8EdZLbEhCiTX3Z!V)uLj znW5sJBj_K(vtCWmi5E2OQAY^^rT2Y%ym$7R+fmmy`t@^#eFu8X%eNUJ zU9pr^0ax~4T*P)nyvIlTR=IZ&%nGWcm2E{{_WIEC&UOtpcmuJ=vC>EIiAGLaGuTT< zhVQAp6OH90Tcz0@4Hqr8fnp9)B){O!-+$5g_!+14F4kYlTJe_>qiVa6Ja6x{QISTB ztfewf@HO3$!dqK()8~W;(m&X=feqKU!7h5XO|6E>I=t<$FxO-P(`B=i(YX&YFUA|p ziIS%z#2YWsOz*ajMB}|HfS$)~uV6Pp59(9Y%=Xmke(`^E6Hj$R50}b`+YtY65u3se z2630;|M4tgi$=g@pL{*6``=27{}oa;{_X76lWO~Q+~A|3+vsV*rfwZ4Y*G6AQkYvs z)SR3uqnKL2noYDKTk<$*)H$VZh9O;{AFQwzr!b|!UmE^GnIhcaO~9`^#DR}PxJ+p! zp=8Fx#mI0MV@Gu((-owE6MU;#Pm!9jjmt`I4xE^=(5zfIS(5)NG`D0UEkp_3`w^HlNJvSAa^PH=BVmzg>D{v zMA3e-z|X{HnJx4qq;>wt(j|nI$9_qe;8@;YQPcp%-1b%Dw3jGYaB=3PPS z-w%ii<1y~3WR^K{;5gmuw2RVIHLGSY6XE0krJDGwWqE#SnYu#%#IM9`x4NI_vCcY9 zCcK=BRM-Y$<>crBV4cp)=>Q}KHPiWjw9s=zkwVhVI9O9jAO>r7U5A!A2mZ^0r{HIm>5Z@rc;gYLP-zzA= z>$=o-WN9ln?a6L4Y!WxJy-_q;|IS+n%O>}IbyRX=xz<+0CePr(NgiPy!y|YAel^|G zN%bsj+L`+%~}v z)8ze$SmHxyq04ZxG$tDCD5XlZ{EA4g@95>{BU@kY!<~byUOXJV5{&6~MUv-26<5RM z0j8t!&g0=J6@PbF1u`t~6&~=L__(?oyR@F;VGfl4i5ZaL3dB4oTsH9UV(d6;c=1mf zwt{B?Z>~@{u6Dn|vo|yZ9ALBZ|4%_)HNmBmK3h?E4sIEuHmo<>&Er|{Zw8s?YIdm? zkBU`W-=7AqmN|tm;Q6r)IBy6u%wXAhHGs^VRax@rgn_vz*5;p9JI>; z)-zl6Xi(+fZ#ZATMxJ>gG@%jk5v!`<)7ii);t`psvLHFIz(OUR*RQYpd+`Bd;_!_` z5w3K?uE>UIBzPLd$7;@^p&8$;Of)*s)q0SJBJp#ltscpy(ugw&%zeBug zi6~|T?E)Qhho@nN8;}^71-1CwhYow&s2P(i%+u^YhEZsaD`~Pvl(qK5Qp9hFa(C{> zX$b~9`C$rFef7ZJmx=f=PxVVn)DNKxH-+6B$drJ6!bw}jo|>k)1XpsDTB(qyPybrA zGd;VLYzjE(T9Q=)rc6~Qm~WrjS21yOn@LT=u`absf?>>Esq6TUriPYN^T*m^y?f87WVrFz|U%RU|Mj(<&=YFtPONHpD2G zwOTY_0PpiG{Xp!)l!Ba5@w|*_iBG55+LPyo3Z9%OpRP%Cm@{v%8&V5!KlFMlT+^Vy z^xSN@ME2#G1*{(t0Sy7pAaXRHz}yyebbXkKo;LkAxU8YD<+Lf@7&$twOr5u1g2Sw5 zFfz|UA9l+*UfJ``{BQ0r_Ix+*AALw6h+b$69%Tl%H{mX?!QP5({OUrF9Mo9_4qH4W z!xueB_p-ZE%O5sP|Cc0&HL%wA+{V?=mU-BH z%kxNyX&f-zvf(tIuytK1IMO-$pTz4A>2u1v=&xZS_k~7mGgu}5nX6hide_WB57fOk2}%&p?AX>Idf71@~%#=#SZq=E3~1yZBj5RVDBgT?diUP z=h@oE%Sl`Af#w6}NIWah?=#fQ>p6*|^6zrFDZ^ePPYwruxcG^~V-zZqz7{Pon+Mkk> zj7*1UEpU(P_`Rp~mdkVPa>FEZy~A}8C6|%nvt`=<4wjiyzP+tb;z8CBP8NF$w_S-nc% z41@Ly5f9wozESvoet8K5Pqo(c0J&x$b*uELe<|(NQZNMvk1z3anxTCXXAltEau)*|?L@|pi9!fvgR zt7Pw3_taUo_DBG~v#5iH9aPJUQD7h6X zDE4~`xr^Mp>>EFBoKcQRjnv$EZQA4h?Pg!pGx{7 zhl#LYV|>PSae>7*RZYDsoXcpaX}D`#dQZzvcQdmgLa?2Y{?7Z|ivCwOMoQLL_qJy% zy;5r7D_l}mR@NAuh4K6ENP2R%Ff+q2qjX=8?{-I}&34-7rO#yi4|7!9)?P)DI~UhQ zjqMxtt|NhIp6Ra;^Ya?xCK;!(75ab10dmv=zXf9MG7g>JtR7nF-PK2FC9-V`ah#WH zT0?H(A~IQ>E-x6Zg16oS<9vQ@B|43HZ-3&IYqrZnyO+ZLPA~FzOGRYwbpzzHp3-OQ zH=O5VnxrJsurBJac?^n&n-SCcuz|%#?EYTKFIp#r#4VGl5e~6|#SW}Uzm$8FvCxC~ z5!+;XqT^%h(9^`EP43oWU=6}IgOBv(pgp2lfe;HKBQ8zzy1-Lyb8b4J6*|DVKk%E2 z$Xl4pbVwkN{o$99;Q&r=0eTpU<$Os|A-*DShCjUjTz#7`KW6&JTzjn3r2q7>-ov2L z^$s!Q8vQ_$=KuUAVR}jNC{zoww$vfdP!hb^k39nVTb0hiCX45>gV%jS~p;jZAxqAo$qVGUUJOf#VMeT^dvg4&^Llm-fr`flz8nAqur)L|{G!-6Xzdlwg# zOE@c2WERXdkgQJVg6Ama7kkV*d<7G zZgqc-k)9p`CGm`Y8$jH7g(+IA3#YdLn)0Pd6u>MyC>2kV)f*C#VN`7z@qSb^d7uvA z!Iuoer&oY~f4F;;VbM!w1+2FCyD4tnM!1jr3O&`o9kn3qPHuF~OXJ58~ z4T5+b=Q%5;4gw!KW`f;kFZEG?ZL&cNxHx4=&FhvBYRf+z^tC1aL?uc|{0*~i_(Olk zAK+j+Lb*vJN6%k!y%D_5`9gemYm8{ZC& zzR2+}ncn+)M|zaYYTv-t;L05nh;p*h@U*Eyt5aC#*vQ~6RDP^s+ND(mG`~bz=cll7 zw0p9y!xn%onIxHJni9^eY@p!Q#j!?kFW~j*kLpgO92;E5%z%2L7m;p6i z9uc;eL=;KLJF!J0JB14|E9!IAk8(@=S;>^fR=b|sbO-$COVeR!O-Wc6B+;UtHSy%M zTRb}8Xh18t+cx$kJ{mDbN&G-onKe9^iCsw!4N1!JM!Lu7*3sN;Q&&Y$ zL$2@!9qFt~R)B1@-}p?u!n2yMXKvGcPUZRJvP{B*&Y!(u{jH>~eu>K+{b^I3P~3GJ zUlk*U&DaZHRl)ZThj{kf*HInfZ{tLtFDGXcTJWd~YGpCLpI0u`x z3i}nXm{SC`FzTBqL;5^1^f=hvQCb>~ENrw|XEZ>AVw4>vVJ79=@piRzNkFa{MO$%P zm{~x5?F2WM=H^99OF=QvVDCz;r5p7#e-RqZu_BMtEjD8x4zk*I;pjek8qamGjg^md zT28APYKNN)Y9Sny$IG?MJ(!$_gqaMoCINPrsq=vqQ3T}kysWA-*q}?u37cC!V&qbf z{|V5SwINX~`zh``$Hhyd{{q9c_*xvjw$|>{apy=xT&scK<#&<`Ke{ANjd|$!+&{ea zduAce=VqW~r03JJ^z7u{7Lj6KwGs_oJ5`8e8V`0DcMNVy7oM3NbpSoX;@x44L_a_& zE+kK-L+NI$a3~x4Y?Y9aLR3bB`j5nq^~CSeVtHB{zfFgb!9G6 zVqgihUWgiiPZb? z6EU$;F`Fh$z#AMW`1Yt~S#=)B^Nv7HR(;{R<~3=|b7T1#Z#>#&yvQf(0CyjDDozP= z<19!%Ht%!p&!hbqJ$15KiF3 zDO6t0%e3*G%A9l2tIA))>WfE^lY!Gi==!}nLP71fRKAm;+0x8AdOPtS&Tv8 zc?~y^+%^ss{t{y$>XjfRKD>V^_`GEiXxw%D9)~XXwpNY%!V#Ft%4``$e5Xr3!wLniDH;HfPWbavNAVlFFd7e?3f+Hkrvq z@xtS5+%k9hUtiuqlo0;mgdvc4FY!v5CVYd3e2mwr=Zaa=ep)e4+?PR%P07tEH?&Mc z$K{Yni%z}nW-BpLE_Qel<9($V7Y+V)^ZUo*ymxaJqM7x#SYs9e`x6G;qLi5pNa>{! z#K)MD!6O=JIU63kSaiUimR6$=T*dMN9ZW~w*i;lU;FH74y&!Voz` zjb#_T2BJpoje(8kiSK=HD;-lehBw5X&JqqzuR$etOt zwzEqj?efH!sm%V69*xIp=8OEUH7&KCIA{#kLq5am3v@b$>d(T1TbD<7KLYcM$r4Fzw*8fTfw$kF?3ZW}0i<2t^N;~Z5s`S3gD zK|;wQR2BS+Vy2&~xjL=pIovG1S;o%pTyfUI6fJU+_Tr$zj&wZ(ryX2|q^Go?@gIyW z7rmWCNTozYe!9WpMKS_k>KMz=Er$OajTD3+gbcpfyC=>_Z*P_?(nb8oE}T}IM;E*1 zz|0&SWyo)BIhhn&t0q9I#KG7t?GvKjy5Iw z2Ug7Tqm#`lq{Y|K5e(1ocD3bb++bk?3XA7yTpUuM)N17}pTWu-!3CBoo0E9ze0hb~ zJ0TFh+ONvjnMRv<*J18JI<*RW9c}QTY5dm4oe2M@$65x{{||mZfxpwh4rW+FW>`gr zaxVcNzBUFU=0vacBCCq(^6N%e^64yGfqiwoD5CgVB$BG%w-f)FVGHSFWh~@*uYbq4BAjB}m*kM4hBW0_glo_@-Bm+Bwt|HeN? zrcmFbKU}B&ZGz=x?7;O}_`8Ttp#5&`ICA}NkKAN=44<=ShnmH+(D|9qLQq;%NQ2)w>W z;I#yj5pG7k`OH2qP0XGqIeTUYev$9~2eY2R(Fp$G5&Cw9rkb(>{kXiCOo zB9#r)uG&mwgQq1<=5xb&8vuDS6|S z=L2wq&|-bEC{QaGu3LsElgMy|1@WmNu|uDo0W3APx6s~+wm@V3vSis>vZ*sDIR|OS zlpvF9mvn!(L{Lw*Evl2kR7^56Q5JBA;HyYEbn86rn>0I_31?+G8I>7YC-p-wE<&@u zx}r_w@Q6&IB<&_!dhwO-5DSZRqt7o7AAFdsa~0rlOiW~D%K+GRdMzJ0epE(BW~8w> zKus{TX%_Y>qNwH^+10LI1f%>ir{h6yBXiVJSID^msM+Nv*%xkzr#f)S;7U^tkmC$HE%GG3f+iD3G0QXU{uA=65_Smp7rg5^i z9_m0HWbGDFUu=e|J`)iM4tnK_KmDq>15n{nzW$^C^?X|=eaHy?RoAFaCPxCd&W zr@CZcUrd&6TqRioAd7ub0v&$o>a)nQny`H3lb6d<+kWw_Y)7%&N8R!{z(osNxxB_@H8Wa{4z+~O9@pp!G zIZ#naw+SD2n3uV^c~})WaRQcWYAewL;eIKR()&}etK*V^S)}8Uro1MO_Qu2<9TDb; zvZEtmY-aZ7Sa|~I3ii!fo&<}?L8l83Wm-jHJt_g?g|T`W{?&}IW`q0s`s9KOE`V7# z{j9T0Oiaih{J|f{`t|FLY)T0jKWm6ePn1UBB_Dy;5=ef@Z~hfHV1lIWz4zWLx7~Kz zY0qvQ0Wh9_-g@afen@OJtE2|d7E4!B+j~$t7uU+ceM2%5@0Yq(kBp7YNkhX@*>m8C zxGTIej7U+EMYCrxB1QV$HaFJDtv`H_l^j5ZkW-O`x6JP+3rJ7k%4Qix{W##M5VyrG zE*AO&2aZYr>hI9l6iX(6&I{cdrRG_buI<#RW>|0uEE}zrv=rIquo-I71r}s=PRanr z#+GCGrlm?wS9~0kM|StgRIpk)t9%kCwY#%#K;~JDtLq!&%JnN{*CTsTf?fuotU_RF zMuNj}dh7y(>ueIAn?@uGK9U_lAh#3U_e_H zI-NzBS6L%ijwl6J&RR|uIYZ}yLV=7l*O}wEW`dd#F=$J{W--i3gbG2SQ6zi?1v zzVMG8!1%?UC*GlZ7~AukEy~Ey^B`_ZU{2@1S`)o;<3c#D!Bx@oaDjR)dMbUaw!}Dx zzNR)u4`MvlV=Ajf_v6BOdSP2&yC}Qn#5r}Jh4-;9I~vF3xdwE5n(CXLdSB_+$saVqLz@jF=)P0}+tE9In& zn*eInwY5^r6IsSU?A$fVB3waanc`ZVT{Z{wY!AI|%@Sc1jgtj-`4RaEh7sf9VCu|5 z8CK04Z7HDrZgWgv!$|R4qw-ia03lIp!tA^Dl8N~ZwdK_^{zR7y_k`uU`%#Ey zQLm{;N)W1ZkzT>HTt?6;CvKwj!^0!8WVuf=WQYCkG6_NP_5oIFp;x=qDFF0=aAo)Y zqmtnoV>n9z`WlbYtG6-Wh+t7PYnN9B+I@CwUe8CWdKnq z%cK$N%F4+;U~m}Q1pS2zJcgx)qIl5ALbTZe5+stJLPV^YTIrkGTjfBeTn_GfLVnQ@ zK~OCqbD?e-gUaj(w8_};wEXmdLkgDT(sek29n46$-79aps6!TImWUIfv$0^gr03=g zComQrrj|U4^7Iy3Gr7nrq@)JtDIVc7ujIOM40W-f%N2dB`Hix7645mawa~amLdmfN zP^BK5>6Ms?FDXvi({}c>U)jC7fY=fKA9%{TFDHbZk*z%$}1&$<%X>|3n<$g z*}v26{mkbZ+d3%$X_EbqZF}YmrN7RK5diiA9<@KaQ@lQyQwkJ;r6$0oTXhszBeG;m zt$;?m17SpfrenC;`Cp!yxNfO93A=YSO?kUYWkLPaY#{kFbb)IY`Pd1nYn-{jHJZ;T%8?WoJ zITl%Bmm6Trb7o$%1K@pLpDb&r!=6_Wq^=~$i*O|CT#hxkP{;z7VO`N_z=Y_eZfdvl z=tg*a9&|<=lB=tcS=NIov@e38c?28*0fe$$^hYcwC>@F?S*N*vo>gFxCE%_1r2z#C zsrdqHN`cmXC$8K$N$H=Ld;~O0GEX3*{I>`%e&7QiFa&819z6JxU&E{IfS>%_JMdEp z%*4}xHU)Od09jx`C^a#gRW<}a_@Cr#q@|%!yy)zBJj_kzLmt1=iEo&K*^$NZ%J8>> z$V4K=F`3iU6jb7?xe2bRM2k|`wN0o_$_)OSzDFPD=qRF~(Fn9LH}rRr*A4K=KRR8E zDRW|$a1N&6NgH6C_n*9-nBx}iLw}siM+co(Z1{l=k}2gLd{u{264<6_B1?cPB5Aa0 z3fP&2;s>lAwn_iCpktFmTgU*!`Z}jlkySM;z+HQG0+H z((%pVo9jLwdf%VOpS=I0Ctq&q@hL{21d>lNC@;ZNci(-tTzB1d&v_(kGSV$E-wJ?> zSBlVF+n3hM;Vz{Jx68zI1jXWrTz1vvs2mRfb_x>D9h9M-X=+fekiNk#>FVmkaAf7$ zx4c`v{ZF?Vk#92qCx=q?mMxdcTx>?1P^x|H9nfW`CE#_)=pcaGjKCElTT>Y847IS! z*RGKS=bsv%mn=Xs1bCi5dQ|dcm`(H>_Un?Yi)a>ln?~SRw6n~9E6#-0LFr~<1{R@H zTTFIYx~Q5yxfOEgu|rUd8@TScoU>^&*~ua3;1TH?p?7a$77!VhwHK`w3l2i{;uTW6 zF)4BCtt-j8+PE+C484jE9+fE`Yt)S^koKwuX#!-H6DWA(!Czza{L&Eclj)w6=9YRg z0%?g+j~t`#b5j5?*sw@yoDDKEK8Pywv}~Y0ap#f4lAH}X`Vp5E#2QP|lQ}QJa*LE<#Ow2u5{4H3;7(Mo>&9iWJ0b&pVewa*sZk!5 z-@o}n>cHCNjc=lL(|BAC-@Qw=?V4oY^JK2)30YT0kLd=#lsB}Ai{l9DemcXlh6 zpK<_R2T$6yewOve4LDSKbCxWp(u^${fW%xWZCKXbkaJ;73-i6JCL#z&~{wT;+y0&JhG#jkGvSz>x8n*&m+Uq%CU1@gO=XFue za-J++)WN#5uy$&#uyVKYp9S?0uGwZB2efWDwVFdRDAPE;PVUXjdaG-$A1buPTqjX{ zg-|RmQQ90Tu+9-sv*;MAqtAXu0D)sHINltewVUB0;7&cWk->iIrb@qFmJ#^XuYM&T z`p|~}ZVCCuH@;!CY0~eNG6gQb{BluHe8UYl80XfW$7lN%M}C3#N?m zV`sj5NQ{`x%mD(Y7FZo+d`~@Pw5SAJY8ir;`LCh;Kox))pjlJj$XKGmL)k%sgvrp* zAlZ2ppv|(cLyA6>VMqWeTS5VHi~v|Fjjw2&5nv8ZR)DNNt_H)532e-Fy(*pePyGhK z^Z(PE@vIuBr#20OVG)e~irNu|xtG>^tBfxDU`hMCWh@9Uc{J3xz}Y+G>{;&ZW8pUIc4%B)_!8qCntF zTW}^+92TEcgK8R3VCYKQ3mkzGNPd9>_8hz4b=O_;`Okm;Ird;pN+kV~?AtHl@@5$W zsICCyskqaDJqN|ZBD!itKyG-`IdbH{HsZk09}-?U_kz`O mc>pG+gLXVVNzac- zU6nE5w#9m&3R(bO11+En~#mJDQ}vhD9CdmN7-83XmBb z9g-ZzBS?mz+>@6j7cWH|+b^-PxP0$R{~|2_R)@_e_0?W!Ypx*M>5!v?6H*0LdfWEH z2u?X<>9Sg>MJ4;jw=I!>`tD8{j}Jf{i_6uQtdP}9{PJhN|G&k>IAx%X`?$|CM8qN( zrsJbC^3d*mw2T@Qhqpqu-S-Qb?iwVUjwl(!PR8Y;=Hve@RGgF;-ome0dZ3UmUUr(Gk#RZdu7S0PML{PSA5=i zC~OlQwNG=rR@t;}xvZ>p$;7~@OkynSQ5C;t%L?gO(YN2MYkLQ zNx7hQ(*t`kwdGnb?!n8dcG9*h9EFNMZq%y9F}8(b1wVG&H-}{7X4$Y7`ZW!kACRwT99(^~I-&pL1nW6Z{8(7K^jwAm%{m9aiq~kdhz&O5=)1Tk4hvBC$(yfx zBiUh}6v%3&0Q7ppRy{WFi%|okfI7#yl}$ko&A5Mch&3Rc%ad8c@wahL{;B|hP%Bw* z7eU>{(%3){n){dv2GOPHWBrK%#Pwd($!3k9ZJ!x-mS&t3G;5ICp}_&Y!P54u5zxBe z&d$!4L2vrMakGS?XV(|C~Uw<1Xy8-0MZMZod!&%Qc7_K9Akf!sluC$ z<0*Jw*mTnC^H2Z9=}!JzA2XMkZ~7mOF`hHm#x;=V#+DHFf_13@0nLb80Q$!M%){D% z2OE_&O5xJngP}t}5U{4E(x8*;emqGQT(ih@A3@V6uNiNEx67r0E)pQv0%gX4DqmquJuMYu8Fs6Ik{+HtRg~L6`?QM0*~4 z8LG%AGh0khqf-Sre^ZnE@dw@_cmCH;~x$tdbo&kI0(y&l6v5Kq5@|Niw0s)YaCb0KE@I&@{g@MD8=3 zcX~E1c_QQWk%357+0PfCZOQb7%RK9-OSLAO$|7{uFJ1XM?^RkB;Y;djAkh%5Kptq1q z%9_Pcl>72xC&OfNW@N(>FUr?m*?R6)S>95KNY#KO#t%!;iJ`>N>KzVCZ3Pr%3}G7% zQsWgH<%aVXONyG^DQJ29gNtNfVpg{AIHKl`;jt67!(l@;J5Lb6k#xxPKvIS$A~I~9 zq)#%vow@H?1XOG&Qg81cCgO?636V%Q0$4}KM#UCKNQJgZ7AO6fr{|$W=Ovz{Zf=Tf zFxll`azuiU?-qZvOUxk;8RZr+tvp|TosY}xpg` z)ntP+(P?o4lEcv{sdqF=GD_DdJ z`sZaFf!7-tK6xnb{Exqr%isE5sjO*`Pki~m3YAXJ|z8)o&Ys@Wi@dPN`Kblp zsiYbJz-i@LR3tHph?Zu-bv?DRz`3|)H#OHn>5EdS-2@Oco2~u7wUuF(K&{freLga0 z74=fz&_oS?wLG@xNg0PJrcQc>wtrsgkJqnRD+vO(GqW>LyH$M4DHYT>m$qk*!1?E& zFaP)d{$IKM_S=oR^=I$#{GZvgXOC>%x>Zz>RUJbWfC>+Q+!RzyIW?&&tPc z=4>*Vq5O^VG3K<=k(B{qWxkv6OEq}ourt@eab~{yD!LdUn4@|EzM5JzQko23)=Wm& zO3&FKB3zR*PH5)55N7IUQU zC}K+yCaaJ^X$N?&k;hp8ClTVRLC|VscvQxAKPju$tz$CbC{U+-jAp{MO_lQGV^7fk zI3gP_T|xFUBl{0`$(Vw37FP?CSD;*I#GayGCUstxleH9p))WF$)BF0MR8#wf(q#tq zWo(YIX_QWdC2|028-EG&Lc4>R)4tIPIEHQC_}S;J=VjrRKW5O$G*s$v&3>DtqP9cU>sS4I#Pn-+n|- zU!Rzx78#$Pm1!c19_TWm)QFL}T~gM}B8)1S20UrSclt{~=oy?J4 z%|kmIolel&q990%#_EbWwkpbn>(Fy)3OUCOqTWW?p|qGwn%ml4ZQCf=GxsU+JQjUaf26E=;5 z;(#WUB!XPPFOFx()JI&Lze4Izls>PM&&D|+5AB_VSIu>lqB6JgO1q>sLN;vwZ$L4*(*}vUKfv^4I_VOJkP|ecZ#8JoqXDhz&hc zVOtmcfBpPNPCnFY`mq&FgqjsAn8&oITZXNDop0($YE7)4wLyiY3ItHep0(OpvlAwb^0)avGMl}yq=Fs}vwD=LFOd~j` zlWpkuCK-`()+#668x9=p<(*22qS1a3K^v;e}r zo7T&H5ATp8y%Q|yHhSgO%b}x(0qAb&T+|?Cto16K=>fo>t=r%I-QO8~fnUh|X@J8; zc5?ah$zHAphxizX|cyavwsg=nvFY z2MiHW^;z=(*y*V$$&qc=x_uMDBPVP~{S3{h(rcnvStnL(hSm?8i+txfzP5eSy8H9j zuLUT#%YzT?K)(l8oz@>?6Er)mtQBRYSOCRZqM`Fr=e}mLjl>W>A|4pm{**J!W3?4( zvy`2pRA<#G&;XfUOB#TlMx?czgh?3&LH$e{b|ah5pb0aKpG*>mEO33TAAsxu(DNR~ zMpev}VX#n|fxFIiYXFW5GK6sTe0qI)P5PtPtToiCrcZp1-$?VY)~I)`<5JpQ`VlCB z-M8#LeU;(&O`D$k-| zT8P_a}dX8wDI;@Agj*GVy zx-uZwx0uZ1@mV7huGPNAiht4_S^-}Qr;+-UILb~-mo*Z$OJj7J$Xp@}24y=RN5G2v zDcgLm{Pckx(g}%Lf$AgQ|25)QYo!^L=p+3-P_aX@YUO$ZxM41-C99OVXi(~C&ot51 z1@)XP@e*ib$0j9)b6^+-ve~g{L1rUqS$SyyTA5h_Q1P@+asu^hFAAMzRkrrlNpF82 z&efPyGd>Xn#(Zufkx+4pw93e%COu^<%H&i^{8kTH=t>z61;v3{_6z`iRYMIK-;i8~ zx@n9%-mL$3G&2P`#(;vi#1_SGj;{XhddF zy>1HBQ(saIwR{@dpA!cMqelk35(lo?i_y2tN&m5w^bbtSRo8BjigU^g!8BKU2U&%% z3?A4gZGm#bu&&v#9BuKS3*XpCu?nGo&_aC&|1?~TB3FASdUSgoImlEJ{{0*ZEZE`a98G#J{}bSI~ORvi{sPl6N=Bwg(@U;phW0m!RG#8J3BdS!QCO-Xs(b zvkQa3L^Vg{u>*VQ-HS+8-)_~j7s$&Jt7N)o%E%%&1EzW;LHiybvvlh_H_Ky>@08VT zHL~&#E|Ks3s*Bo+0CliB>0{Dl_ehKuNaeMQWwLifa^5Up(Mx~bjKqc_Mua^Xn#5oL zhlyCL`eja?529yTGL^V5S_#D@hyGj^mZcqa2w>3?sI5ty?u@Kjwu~v6l^^~5VOefj zC3oC+xBTh<`QP&7Ap!zWke47P_4wg~vSiB|=*MI_0qPOdYsY#IOB3qa6TyBlV>m{R zOiP9cdjWxzmJQ2f?*!Krq;`Xde<7cgT6#d=aOp;-FS zhhT?z#M8dPIL7E4y_y+|IGOV0L@r;6EvxNS_#cAJkeE%-_;z!;oGvmXj-RHIZ!h!(JYxxddnrm0bPd7s$ zmft5pvHn&e8U>(syRvWSj9}4Bu&P))Sf9!eu_>o~2Gg8v zIN@XC6EaKvZ5rFCRA()vv0;as$+l{}qK_<*W}P{t;aF@s{l5SJKmbWZK~&~&ioNHo zl5wbwT|ND*$7Exawuw^?e&<{wwb)n9WZ8M2o>Q|wy18IVWdOh`wp0KF+ly{_{ISQ# z{>N7kBd&@lcK0-;Bt)-6ayaNh0k>;~> z`;Y(l54q)*Ta3>}Gqzg8tKToBHJ{if=pp=@KvchTT4Su+u3fuM?E13*{)@l(3xh48 zb;xJ))(jO`gG`HCnv7chk+BH_1dmOaWNMDY0{-Ntbt~oEb*qh}MFH@q37Hgi^HX!w z2=h5=yUqGaTCV|=Gq6iCus*CLKvY~S#oWsQtTiy^bCWz^&MO>F5%|jUSXJUN7i0-w zcTiGMR#q-sH*b^&9(;s=5;Cu-TVrqx5UsTCEHq>jZ_q%TiGJXi0RxO0O40<)3=qov zFOrewLkpnsTMRZKi6To?fh3yVrO!TSJ?GXd3r z6p7Zez}M846B$@4iyF$Mj@st!|NW@c(|37v@Q`HUS^4#^w#yY4UW$OuZrQYUo!oc( z&m@FlYF!Jx+u@L9C}dk5W%O`9OpQpn96vGw=+Dy9h`z;UD9E#;l9&i$+~^a$ezo*G zc$^v&DAQXvBcdXR`HV^mN?D$<#WEVI-CAE^A`)lfJzg&v1IARv8m6MuT?7ZEg{;6L zve7B1+zujhQR=@`)JjQdl|=eYI8PR+FfKB=afDB*s~i%?(8OnH1I6(+Rk~#D`&Y`{ zPYhz9i8i_lvS#yE`P1LOLVSe~K?5i;jMI0&{SR{EJANOUXtj)t4U>sT%A6-jjdmsV zClNV5FeEo!a~b+(NcoeosHPS>j&N22$0tcPJ2(a8rLETB=m?G%yz4@g%z}^J%X}}$ zJm+zU@&M^%6bfEN9nneJF&&#?v1k+v_cc4CbiG=vc2@d_QqslyDycqu2Pu$kIT3VN^ zq;AqlQY|IDNA^k|KT~6a&wk^pSzJuTD z;AwVN9ZA)*umFV1$+DI$v{xcyq*UoBtRNNL(oBuk)M;B6H^(#7qzT@ZA!g}?9<3B@ z1(B8h0PSDg23<3N5SACJp_#Q!RjT9RdDaIqJ2;lCh|Fu5*V;z^}OSOzfgWG54rV7tT>9Fw_L5bh{ zz2AHKCTGzyQAOx?z3W|a-+lM-acL;xncH9e)nADM)g3!_JoAO;`0LV3FO@HT@rwq~ ze(!tVYh(eHy00|p=h~Dp(9qBzmtA(5eD$keeXd>4iq{emG*POmWtKK$)0ENm^=kV` zpetW?4M=$iPSpdRt@y1L7;V~VB~J&HD6LB;rl%->o0ANT#sK!BjBvj=XeWDz%1uVbFKpH zlQOhgQm`%b+o1~hO8aJp8a^*miti0`T)Ek7x{rx#SG9X50X$`!xHRjm%oKcBZI@=| zO`&jvIZ7Z)F*Mhs>f8<+ZB{u}URzf|-PU$g8CZRG1Z}NWJ9C@*XKaGDq(g>}03Lz2 zwAu!IaIP~)_T1=G&il1X3J&j6-)a1H{5`dycKmBDzqGx`5l}34&D-9zQfZ{oN5>^8vrXQ z_oF^-CN$-xCg`Gz&zBK;^=co+P=u8Y5rW!=HfoR*%;aRCcR)t^Cz<2{$p+GQ4L>0)ul>oV6nm)hN?p0HWEC;7y3u7UdF|iZOxN4{%weWrn*B zrDJN4qlkRu=_}`;MN)=|=>)VTFfODh#e;EE!L8g(LK>Q!vVOx#fEmJA2aZYk&@j|& zvve+5B}We&VDYa+;2dy+oQVOf6uCP_hq8=e})X>rS7-w?4etv8E_ zCI-q8sYzzIHYWRY{}`>H5HcZJ7^c3=fdyT&q=hJ3SYq@Ob|T3?qw1nW{f0*9hm515 zVh1W^Z^a0;b#zE>Iw~Qu&STV+sEmDNc8>Rkak^-mMHH_N`mxnnBVLzJiUb_&j)F7; z`0WUx#G!*5K$X739#<1o7mSiUE*;cqU$m@EZrgTH<}lDU+6|1N1U>DsN^MyafYmKm zT=HgGWLc!T2B8;d&C&Tene(F{IyDa{O##ZwP{CX+d+DLPgzN06UoN9_WYzhv$nm3$ zbG0AEZWfa&jJ+L4DU-7rl=-OAz@h7;xc72ISaN(BmA)M%;E};Ndk~(;Q?mj;TO9C? z8mQ%7w1rGI1~Y@ftf&acERNVKvDvxaleY{J(bvLO0>h{Ol)UAY%;D$14Rs$D{duiz zN^^6fka{MYQHy#e2RKXKtBz}e;H364)+~?$U2T7)^UTFK`w6fHP;RvuYSr}S#fJLJ z-DFKth&EEUOAxmVT5=8aStr?F79wbrIF{-pli~FuVAbAO2XiKXSR>+y%6!5JF7T5| z5`NWyDfZtAIJT&IEVfbKpT#jQC&;+8qfzEV<;JnJW|^du;XG?ry2`J@hP0F6M}ct+ z=P86u&n06xxv0nzfR0g5se$O6_7!DaOcR(*z>G4RCb-7D93cqX)L0{J1i{@14`JSo zb1?+0Q`|4v;|$r}1VN(=86!K_?;`W;;765pG443^M6>yOHW-!ZV2tOXb5>NLX)(pc zvIiiez<9>vMHv|h^VG0h{)V$#TSBw1CuetRYimU*)}Q?3C*|II?>+78(JY4oVQ=MFcb+RdPOTtK(7j18F3W=x&s7SjxW1(aho*OViT}cmZHb72{|&u z+9qmrVq8Y}$#A{O9JE=;Si6<(jBqR8%lg_H&XJQKVr;72qvCp1uuD|aKq=B;u0{26 zR69mrQ#Eh15j;?k&iL@Wj-%Smg@Dh(xMWETE&R=Jl(rW<0`rt}#h)rWS=zb$Ip43J zpDLlhB!TniJkzWH#UK6XNAl)3zu9Ox_I$3^8V1N#6DhXPZs`LXJLC(KCV*B{sxjQw za)+#_+aQgVwX*%5d*z)40OB5`JM96BYS1vo)q=RKKbRg{bC|w<)nrx2e2#HnDhW{H5eAM&u_Zw zG9ybHM!njFfteW{Bf>mpEXq{AGWnFgOwJh3AdbqkNqaxy)LeSi8fu{r(xiU9Oca>l z%bKX0jmm%BaiV<>D(0<&0`TDo|CHk&wWuq3r zP*+kBsG=TxMv|i#H7G$gqLE9M)ycItT!Qn3;`dBQzV+kV$*@<_hRQ~-WI9NbO>SxH zFcuIqJ)%={C{EiE)tZ};GH688(InB*773A6vv>gAjf-)zU;)5Ba9CYZ(Xm!M80a!2 z|0jlrX-@?;&qCAtLG^zbU-EryEIu|AI6%X=m|P z1bNgpV`I-?D*@w$!80<$tSLIjG&-M#_NahHg<~43Rk6Q}*LV)3m}UND$-J6qgW|(> zxU3EXlxq7TZIEJ8M-iuT9x*F*IJFytZuN9@aIp57akiBXscl)5da2Z7)z`3U#)Wmj z0v%DEnk-pZrEn{%)WE8nZUmJCS_`Zf5!MbbKwR6$q_`Ghj@U4`L8N7slI`T2Wn`)6 zqb8i;S*`_b2LinrvO(J4RGn#cbcsfjCALb8tnW;8RC-3{pm$drj9uHP`Sr4;@Sz5Kdk8e80mgAk1DK0^H_Zg1kZ#4taTp?6+Pm9~ zO(;;ZmWB$flxUrM8Wd7he{D(G)YL$*NISgX_tn%JWpEkl|#s zx+>JM@!dllFBAd1GuJEt<0eF3mo8aMrWyJ&Hbq7HR6SoAFJ^Y~paD=Zk6(aP0FgMC!QOVY!JE#*~mPT+SgDk z{k_z$lBsR41L&bVJvWgc8<+t^&4|5;woHBFGB@lXO63(t6;ZExC>BJ?98AbEYOi+f z?~+!8pjI?m<%*445RxW)|mzY&BH1 zSPD8Dlo1D%v(aNvq*+kv0=43yEl~=ZSdKozwg$4vb0H&3>gD=oaFnzY)A-?|vbedK z9>-JCu(*YcgUB&5#NP5M0t?ePS6&$&oRsFeN~HM*EZKrQUlY+xE}7ssiD2b7#VY=Lw76<$dd8~ z@$i{s;>YQ&40t9h;zh6wW({o&5Ot|rB6mLcgmkpE8ft~b6pkI9R?$y@fCVL7UJC`t zB*C6Z@j=aw;1G7>hORTS`7@>pwY$@FM_E+t4!I3{kKQM)?3 zIG|YorYfJSlNyczRMm0Ck(?sqY-jzkSkq81DWaieB7QD*aRD^u0nH12g`r5>$SUEG zmzUKWEF4w(*P34i&>6DV3UahdfE6ddo;9gx&B^HKs1bP1gN6%$0%Fj{rggvyRx=d; zLUMuuS&$%E^C{@1gCirbtQ^up4YZfisT|I_ii!cLIX?GI@W2g|XcpDl*nP~jsS4h@FTOMmGETD6RT*a7YL3HL#tO7<3NBpC>*Qap``)*2pM2&spD`%U zcieHuGwxhJzxTcGee$I*eaU!UY1hwpQx)tV{_uxJIlw{|f8m|q`qsA$V0_(m*NOJ? zJk_S{n|}WDpP%Zbv+g;~5DyH}mI&<>Vg$?w<~s4GqVkiOM~#hba^!d)!68~3`B0hG zZ>$EYnETqt+=ZWM$G>sHY*c24+6q+?*DP$7gDaCqGsG$EK?)n8EW3fBar6S_gj^7F zk#lmYOo0*#fTOZX^w?Rzw}!#;T$GidmIr6j29;Us`Zd6oB6t-j@CD~X66d-cfcNSENfTHn`d!sy1pE~;H=h8g#{klti$I}n zdd(VuQQM^RzqGyh5l|cdryu&5eC|tsDOX;9seI|LzDk{Y)iVy5B9MCc*kSqgL)#2s z{EQc0hkx|+^vM4G`=57xu?~^->;L#)(z$#YG~}RktUd?PtaU`RVsh8dACT;POkz;6 zOi-+}&+gnX6Nf_lBoQ!;itFPrK(7#u$!KrC%wiP00I%45QkJb*VHkt^x9z4T*dpMEM8mHc(`mBW2$RL;CTXVuJq<@Dhw%th1gP~zWQt69 z4BAT$V>CNCO{*gWR;c~30if&5)NVJpq`}iBSq$g#!KfVS8KdQeQ>tkHQAkfxj;j$p zM4=Eh@nfXjh%`)024!q=5~W194AXK$`vFH$;tt~!l~qz>y-7i~(u1}Bn#>~9@>xU| z+UV81cFk%d^ITg~CzHqaO9#gdX!2Y)9({b(;eG)O1a-54Jd(cd4(pge@W*vfN^MOP^yf zH?$E*YhNl;JqIMDia%%J*5LBWCD*;@%+BT1cH3yYEs;YJ%;9qeW}hAv+f z;5Z=;zYFo6phN(_^H4D-VGqqy_gAzeU>iZlCL3c0NUI;I;x;MPByCIRA#>f?-YjW= z6_hEQ>m2pD*iZt&YU?#{tBSB@)=E{ZP6K2WylPEg2F6gD7BwkDxJSV$&MH4yPNI(M zg7bO;wCdkh84Yn>GS_32^8qy>(8`!;pbIfha#b))+f`*4qp>+5a%`q#fM-~ayijUL4ZDD&FAd$*D0RhqR5 zb1hl2d!MHzma1BQeXP^oi9%lN{u%dE_I z4;fYkh6+%X{g_5@E()t~g6F2DsPAWx~}eh7HqT zn0l^cvUyQ8f3ySx2NCgA0HQ!$zmv3~FA&C04N+dF0RXLARwZmp4uLMN!~>HPZ&IAt z9xd0?doVj0Gu?E&)P9z>7dHYbymj?k zFMs;YH{Se~ryrLdoO%S*#@=+(O{aeGG|yUxhYC~@Y@sCGo0*_N>E|#Hs4PA|zu`JGrg^W)Mg}@>L3Qg7Zm@me~PR620_5f0QmbL-f zN@N}ybc~uZ6>$j7lfi}3J(-&lSJ+NnNwqXzRAtb2hM0hmXQaE2IvS#qHqIXe#EuS! z$f7H-4H>PbhDHaW8&{)t=)w>=0_Iaq7cs0pBjNDL=mBF8VCH zpn-d#6R~ixNK8)8aSvp`%V88i`Ff?c%2lgm@b0^wKK(3|v?Xu2`su&_Upy#fhF{uI zzV^Kzmf=r-kZ&?}>?~W50|1bQ`c~O^*>z_$Xa*=I>R;Lx95!nlUhGqNx;<3 z{yrS-YJ!u)Fm#3=gdqi1TsTJ(iZuCFJV+JR(M<(%H2a&!`<i7@BQ_*{on^bkdBTH`Rl*_YgxN?t?_fJG1$3trvX+^_1X)0?k7L_ zi3||T?(OY8<=gaH_U_#)pZ)A-pUzaDa>uje$$r>+BxmvGj9f5x1+`uMNgnCE2;r;o zwWhUY1^_xi?QT)(tIMRJwgTYD_nq!6%I2`rBRmH4!-WqvO;%L{MRr80MoAz<@grxT zv`>+hHcA(me`-wYTKrRUR{WDTsT9m)_q%TXdyMu00z%S{*l ztphfQ6SktyU5(wK^9;w21NgH{OcX5e?Fhs%x63_##KHi?1OleUrwA$_CPv0w1qL}C z*I{QI&?LfV3~OnJQRm7%Vpz2HT5X2fm(up)MqqKr66svLRJwb+Ufcm*hwu8{_rCYM z^%+hKkF^S5R>T^_1+feB}U zdg4|p1=K1B`)G8`iIEtdiO?IE+E+koFUN%hOw@_jtYjsicOBTZUFzJ>i>=it`1w&s zw#e82`M>1iE$7Ld+Ydm;ZkBVdzD!coETpML{pru|lN!%rYCuuOgL>`vRLJpTgM^)aVg;E4lBYhTvuzRgKz~T;lyEgD6%P5jZ~%&D$@(fY0QxM4-`(U~Ibgbt9Hl#{2Bz!x0+{=Vc?A_g_A^ zQ@n_PmEmw%(?ROBr>T3+Nm8v8Ai0V*TvNlnQrprlaST;sQ>_u)hyXxUZ7``fjX`dv z<~l;HwgV+;q-Vrl(I6Y3Xp9Z)pe{H>gtP<4%SP~oai#7$$)YhB$s_nuBVp+E%L27B z9YG|F=;g%RkkmFd%CY_d`dT_LkYt>Rf;L0D)~52aGxIVrFa`w)z>Ua-1xleGquGQQ zo6^c(sjYhbMzN5QQ3{i8T4OxhP0e}d(wFs;bfG9^hF_i`Ze6-suDj*Ka?cOHB?E}4 zoGsuO*G%f5>Hf|~KX=9hXQocBv@MK)0$YQI4ARuPJl(XO*l=L6QwDg%v?@fj<3Q>b z3I?(HR{DjO^Qo{#!U)p9q}Cx?iVmYLQS0@TzMMs!)kp7MZ4-lUY}7C|HrC6E70YGU zo_!omDar`^uyz(XpE{D-alt}{36u>mW_>2B<0KQ^Oom5wD+*+dQn?gs9YJB<;3s39 zB6t}L#yJ*!N%8(QbA+h{NR#K*pe5^+9$!oUxKbW#h8R#_0AvN^>bz=r0H+q9udnIF z=zZx{RNATjrP=4Fd;kkN8TjW2W0EE5sP$p7*${~}-b!WU$4aL}M*pKjl?;raAT+VJ_5#;lZn{YIE%(rggtCtm{-h0mGSE&hWLTFG7=Nmsum0mKU_6AMRE(k{fZtdoSeK^@O%hg& zCXvkCl=f~nVH>hV<*jJo;dMIMU!qBK(>rdK@BZ*R^6%fgl{Kf5&sT#V|Fh}$YifE@ zZux^3mL>h4Bb+^tv}B~Nu1+>?-1vfS${L@Yl3*$*7V3P)!=tp9pk{&zf9T+Y#(@;* zDD5~!c33Bl))OW9Vk;QdzK$j)Xc{_lUL$8jq)br5vP8mW0YYvJMh+UU-RnR|fFgcV zncVjCU&$q#8fAr#h?|UxR7K^{-FxM(?Z1${PwbZ7v3a@r{A-9_RUtSvD$@fonH@t& zA~PUG7c?{I(z=K<&83X%`M}LL(1(um&+esmwoYs_ad~LR6XL=lQL2b3$u-6c^5C7@ zB+%L*1t@9`vYRv1La9P>0XnmaciD8ICnHegr~lAld=MrtR)q3p^-TDms3IA0C*o5U zy=SP}UNYlDh;7+Sw24AxWOjsOLaDa8p_Wi1wydpPF50?=tac6|pe6G58{RB`^XX5? zqLzBeLN6PenUWvw*olF2(Ly6g?KDCNI5!`7=XHpb9ic8bCI{!z5*dKjg!9o@o|H-q znj2MWh&yu61KTBzA+NAi;h-d?rYtLy(DAZ)H`)0bC@z5aVfrFRBT`_;1a;vDXOBut za|23;J|+~3+GLb9N;x&qi*tuiM3`(hQN!w*CV+cJwp_hRdQlhu?o?xVwiWVcKXA1A_EDK%aX(Z_ySOcc~bJE#5e z0mK6qBUAzH(2|h=aS04RJ;X2jgRR%T z+W;4L{m0j2mQ3F~-`v;grcPTu^QsQz|M&fWUzV!6rWbW9@!1)fgudChX_NS<4J&O6 zBcLj>`ePt9mIQTWPxLgl@ZHhAoCavbR%rHCYek6)6EtL@Pyk8r(!}#xqvv4#i?X&T z9nlcpV%3aSox z7B-WDVRf824Zn1KHU4A>bgzZST1%+wy#~ki{P#0&2^t z`_%870vpZj8sEhiIV{c1&2q^lmpuJ$RZ7-CiK2GxfwZ)Orr}eu&PvA*E6aBvCNus3routlpp1_ld*LA+=cbe)m{rY1g zo*FfsGkCe`FKH-BZ3lZi^&+$9sfWfd{;;sWs!L}`FbxllNE%ABmQARLt6G3GdIl<} zmLR!ep?+8APMVXZ<|ACZoosPlrf1QJL5$0&>*bpDBn#KWS$JF^8Dv-9Hd$sr?bps- zzZO+}AEg`G;*ER2XV&(rhEhIrRqsuMZEgT|$_Nxv7J}VU?HEfO1$+W?iJ?&=yPn~_ zI-aVqzYyutAfqPE@SSyBv41Ku`1FlGH4>#?FKh&Ec=Ov4-`XqftsSR*V>x}{6+o+k z`I&ATHV9(IXL^)Z=KZ(cdaG#5x)-`xZ7iBOCYJVDD-tDAkXxWFQq!81sNSM(3C$wW zJ0ex|s*R!8n-3?caUjaX;?Gx{MG;yD6JI6URfSHMOPfZ$pimT8+%;N;U^4XRP0$9rNjgg|G@qGrjW^!(s|-z7u+2s;&Ha?YBqEWQp2j?K#DSDlZOk(K}c$!+2R z*QFU>WoqmqRR=i+>Js5Kg8%VHc2DSGUglIlFv;HXMz9FwGgqQ^(g%IEsS`EFE%@kbS@#tO|8$|pJr$xmult!Z#3Rgx3X32KX1NbpEO*lph zh=T|o)Yf@9S4^tf&jX0I^8P8wDg7tVK+CIo8n+WvP=#-tE)`*^Z)h;u2_(5sOCAa| zH0(4{M<+EWT0b{Aih~AKr`%C3*Iu*^%J-8pFagL$Ff2^Qx1~OSGlod3s>rHGazSxHH=n{8O99@eq3u*syKSev)dd|+c@yqiEm^$Y#I(5c+$ZO2 z4m?vx;Dvsn&vgIB%PxNgFrFWvcE0ArEDLH$P5Mkfr5FDGE!Vu0KEc9b-5DM8YZw0^#SheqUocB7gu{o4% z(>R3+Zo`z|q+rLm$ttN>k2=CdT_3d@18J2!>x2z@;?_;)@O&0c3(AHxn~TjiRywj$ zHh`qx>x8!Lr{#o?EEPo$@jzs*jDE_fc@2WQw8k1QnBNy zLCM0z((~t7UvR^XjJU2fITlo0bK_&wbz6b2u^AVvOOUSa+&B(=fptn9X|2E3DWusi zeZ9KYA+7XW*1L?N{V^AhN}$RjL(o>uyiPnAUju)D_ZgXMJ&0~k{aX0r?EhC)&Xr2- z!~fd6d9$Gy9itZYi6@?T$!=aVC0fUm zD%<}%eRgMi%WgJp(<=~q4Jc9+q*?$0X$p2!z=od%L`AWne)7i#0{VUGXTgf1(u<&! z1On-8v)OE~(`Wv_=gdyngd`-{B&5tq?%q51o_p>&XYM)kKIeVk=S_pyoE+=kVo6wf zm&F1dwv$gl{=^+_`HJQCp9M>?43udwzQe=`PNPhaMpUkZTHaAgISsJR2?QcD@>vT& z8^AN;n2&Bg@o=o9xy9LU@Jv2YA!&jcUx46LQPM4tbfPp-IT6C6${bmi1X^E8VF4)3 zbqKjJDork0+I~!)Gt?~$c$W*rE5l23k;bmVCb1|7ao=pXW=4&w-%x+C<0p{+2ha%1 zJX@iaG#$8w8s5!f2)HuLDY7()LC&YRGr|* zeq_fVvJc$3a^*_<%U}LtGNs<$QRoop&h2lN;De_2#Zux2ctDCzs0Of z5h%!>A+B*V#2#n9^$moA`V~LY9{a=J9je-~r1ibL2%ltx0zUf@HsV}QiR6HZCB&kD7ZQUUTcARB00Q|}44FS` zYM7UMtcq3GQKKO~0N@28DLd`lqwT1}kF#s9`?}SjDP6rh#!7IRtzNRw!T{-Z2ujM( z@|M)rST_V)0_!REo>We=ps-7;UMBE1{$iR8`vyAPUVhHKn5fIwEXE|HKcHN!9E&;w>8b%gLET7lU*S`17r2M6&5h7`sLE=;CzRy@;~HZ!MMCu9(%2xLLb z-u#%Hvl>8nDJEmgXQbg+w-(0h$un#vz9dz()z&m=g5CRvzgS((1bm$bSY-^cFhGbX z%M6#-@WgAa^YYwXy$`El`%Lz6PQ5L)T=_>RX+ zo@rNp{(IKGc9kuD{$JLzatRu}E`*1iVDoy^Xxrb8Sk)+~=)Cw-3uR}s-@FQBM7ITvS`V-_o9QKf5P1XjtBPN-$k zSORbn;78%{q%jU6><$1D6<#Tc7P&|ZkSs<+BZ=k~0TN52N@609t|kQEdYJ#Qpb_(J zi1|Lug6-6%3B&>}w+M@h9)#OI!XYUVL1_sLA1wB!W5gB3hggiY`mG3K&Rv1lpdTSu zW|#3q57K&qbNs%uxGO>&TFZt2&2OnpBT?ug&?B((8{q_g$?jy4LRl&wXHKnN&^e(A z*CWUbe3t!?!bsH+<}X2b#DT#u!aWNGES#=cxrpZzp@QG$xH^g79xbe^TnDQuASXtv z%>5$?7zuHIvN{1Gp0;-_sjG-jzAcT#>uTgNw2We?~E)L63By2V+`mXV15z!?q&{C zXf|1@$f`m8m-`*n{J^3Y09qUc6yvrO^No~EZUIj8^o6v;Jpt&*admYK?`zo3Jnt=b?wN15&wkv%CIi#~eF@P)rBfQ~z2<9Nha^ zF<)!zF-wjjNMr!xeguPP03if3Zy!I}LUmem%1Wb64CNT?d@BHn&5I~E4Bb3G{z;VWK3{E`*l4W9HiGXBJqvcfi7!ph@qo2~2U= zl1u=s4Y*Y?k)X2l4g1%;g@-w+9 zrJ-OK``{p23Hh&xwmrmtpo2bQvQw$Ao9&^#2TlkubR`yw_2aQu3wrnAr80F2AWjCKRL3?2dNlbqi|Mk9&YF;^GrZpUQ{6^>Y*AGs^qxs*>#5rMB!m7?I zF0hw*P2^i*nXX5BJ8oQ)3|3#>rHoNAU zYqpkcv)2Z&^1J@}>ute;1x`aPBCjpQ^Aw!Kz`Rc3>JAHI4FS zkot+8Uf){@4AE2IyYc#0xbRZ~PjNK3&i%pFg+Dx_3fXS@v?+GLj7hAMX?&vM0r+YW zP8#mMf&Lt*rEe6LNimqEWtL`G?D(=miSk^5=VD}D#Uj@^CCSZ7ZVr{$N++myw8;fS zGOOy~G}oLfq#f&l&)xKy4WM{<2?D|Iz2qvJFm{5y=fjt>a1pg{+;r>kEc?~*kAM8* z9S%L`|9?NZ9Ze@DSOgYV?6EYfQ~Fs{5(v~vlfa<{n4^{FSTFkUqyg^c2lBcw3=9gH za%f6$V}+TgVSIx#)efUUT(xL98d90C0LGXUs%lCBz-jyGPfllGX8@CZthn)`OQHI4 z_VGim;)-Fqjzg%GWq``GS%nLXdzeu)*nIdOhwzW)SK0<4yOP#7qg6mqddgxutrTY2M@OFAc? zFZp|DB~D3ZZ<=C79T7kQ80P1ljC$02(X?uBj zIn2~OkSyd@mHdi3$8+Sv9ngv=KL}|S9r`evsjeFZ$pe1EtPUP^!fC{~?56u!Eu=iv zrwpxeoYh3JA&P}sE8?1L6(Y+=3}y~&Tvdq4HUXI#=V=1BJ3-GU(2Ssj19 zleEU;PdnZGXk8`&l9MklM1XWEk{zyEn7%HgLJz#&J@C@+VTES_Zg`q<-UY)y?%&6N zR{e5GgR1Z=0=oPppz-zFWHc<(5I$=Zbz!j2;xO}T@v^3}7M_I?Jadw^X6Q6xs`bL)in{TkMvm?)-v7<8e^=BMwe24gc;oVE@FX&;2ran+ToKBGAX;W2L@Oc zZ$=>ykE?@t2ebbaTA~5IyHb8TdNTvo){W54`3S!`5JiY?(2Z5%@X;p`{NZOsBDA?v`J&7F@d^e(ZwWvJ>LUo<0O(em^Dn>-R+%B1@kr zoS}Yw0!-jpA}1Ijprj~6!Hu9!ek`myu_%L$!h%UZ!ZTXr>WqEp9W$&du-N=H>un`@ zo6})H04AZ-F~J{T$$s_T$%^3NhaYyp_}Xi)waJqwo5Cp#PttVEnl;OX8j(r#NYelO z&;Q_7c9;XcBH7wp(r@+$KlnlW#V>yGnqYl%_wQ}zG_$V76%=BV>uPUz^G=q!#*mX0(_QnSE7Q#exn@fvtaGyFRB{?( zWJv+{JZRQ&4lKAVg+JTb0}8!&?;n2)Z$!C$>f_hjBac7)suE5-;Z69BSJ?mk?w5A? zMens)Gq5BbL4v=^e*gOs6u_p$<^7jO8m zyAF;y?AYD)s6FmEM;>&<9#^MAg?9Bdss1PO$bU+)6tbpQ>1V}#D0&ovT(^64Z3fO~yu^2ZSq5-4< zl{mLxhK#nD75j>&CbYGqh*!S?0oP|oj3F|D2P&)q_$}g)D0*Y!4%mM<03YvVVboi0MLNf{-_wP(da_BwhS$^A`|rBlF^6OMhnX&SO}FNxO)(C z6ZrHt;PR3p;(`RxZUDNp$fYa_H#LpN6uSbUmAJ{osFf+Rge+}hrY#4Q7e@%i*oP%j zkT|Tqe(UTY6d=!bX4go(h`7$DSjw*<+bJoIkhmrEvq;Utw1~%2Xu17SxHO2F9D?BT z!xtT95%?~a6V;)(3&0OBSVL*jy6JvycjfI*Z=SeAzW(*EZ)oz_XP<4_c4hkb$3MQ+ z;uXjE$}6w5kA3W8b~vPb;v(6XlioWAVgYMXj>rlmj6cD=V-6!kQ;#jf+OfNbYktBf zVTI_$IldE)9UoypeE`a2iW_7*iKU(5bmzga1AzjHEK@RQjAgNsV*Uv8PEM&E z4;wZXgIF6flnE_YZk2>Ejk0jo&$%A>ZoF7m2C#?`CqV>&?!(79o?6TNM*X-aO_$}Lt9Xu1Q)nnPTN?8`6KE^OQKmCKwv=?}hhhaGX~k;K}5b-Z$!z?YX*@a}fo za|@mwF)-eR&Dmu||L})DwD-UN{X1N~Ms^_;dZ5q)Tj>EWns$HV7)z~Ly_H;rONAcT zGkZXUDGPu1TWfpyrHv|Bd&r?y&=zmhB?xahe>st6$%ML|(0DVaOmV=t7!Cz*LO#1O zhz5!MiEt|2NH3bov+*K6?k4f1bzbojJAmF+_UP(b-Nu;+OKkC>p0YE;lK_a11j$g)@{-nxJ_KjZw&tA20@2 zqpAVpSQriT5`RK~7$7eTAIY>*pCj`-;4S5tK(q6=lrSvqklKc_xrn+|5zqMQ%ct6? z)H3UNsly^I_)rp^q80<9ELtr%J{>3y2$;xhdp}AtjehL0$L#2%kKV1VKKI;nO<|z~ z%w_VLPXco?!#9Nm(Ind z49k%3{j|lNnvZ2jGGyIt_~7Dhn7|59Tq?52Fo?M_a_3__h*U0tG_4*L5EkebEY^aN z#vlLD`obf?s!(}}&nf}h5eU8_fKV4Rn1wEpYK8FA&kzbFibbU}*9O$e!cS(%(za^x zLY98$l5JhfIvxZ$E|z*?u5N^6EiKJ2=((l>3&PrBYk~-G{=(HR4!MtOjWvBXt$Mvx z+H!y}R)wg!MV^`gG|MkqG1t=>h){`x-X1TvmKZ!Ytq{dBAFKzc576!g+{-G@gb5nRUzcLj9u7 zD5M?gf%;Jm?!GCLrf%{67hitCR<2&*uIJ6%x0rb!yzuCwk6Lqcv%UT8Z{Ok86@c*$ zcU|Ge{iz3htj_8e&w*pD@D@^`2lj#<=t1kV_%DAZ3|i|(wJmF!Xa{`wqZ^$od_ZFa z38DP`s)WG{GrMmZR~U3A$W9AbEDes=#4+MKyCn(yGwtloR@}!qaY^ zj%8$EH9v(*4|!x(Nf=M`q;l_((2OvO=15bC9C{Q>Rfy8L$`BASgo&od*Jd$ar_IC^ zd+}PIz4QWcxQptoQe=B3v%aO-fgXVzrLq}qy(l(VL!moWj`FMxrDk#@r znKeUMW%alvv_AhbmK42q$r*=R*ZRNPid8Hk#Q})C`!fBPRfOJKy}EjPIpsEmTFM>q zSNT{((wLCnaKKMUO2|hQJ*1BDTvXh_ zFk!L{z!W!ttUZbVp&_)+)pbMxDFVRa{@Ay2HC7~Oj3LSz1#eCewQUim`W5gL;H7Aj z8_Fy_xYz>8R-4o?$BL{SE}>ov4#gGDVCkVTEn-^%e!m}{8E*wqSVrOgP;B#Qd)vwm ztE(N2dG#1KcgW4ok2!cz(HN^{;nfdW8UB3q6QN0AL$Ai&!(1kyN1oj7QRYg?sm_9`M!R!!03s#WMIPq$?_;9zIckdAcpA-5gdx)AW(>= zP{xmkEGEmzE~|*G1puxze+qDnNn;%4h!dtIDGLJ2VHq)60(A!SMFC@JmOP4Ks+#cG zT}?sU73jmn8X_)FuZ<&|-3&C$FF~fbDn>{=1nMkW*>PjW*hFHTmZHIislhr|t|Gk5 zO6T7v4X(H~s;jFpfyCS!ZJ_*CJGwd%qGOH+q8{y_{18)U!HW>SMGU5R_l-3bwx+Bd z;>3Q0S>U`(K>^8DR7t1$Iei%4*KR53NqZS>fi4=%S%^D^@iWFZ7!4{|g$ zG@>MUEYiUu1jk7h0JNwubhk>_Z<2g@rJ4Q5KmK8xW1NFEwaq(zwwQ4TArXN~o^ zJ+SChTEwIv>l4O2Mj{xI48TH9ln;D_(=n+0STrid%=(~N)8T;Wb&Sw;qv}{F3le#xr^V{4Aps0@TXDs5To~X);Jfu> z72|cQbeKLM3hB|NN{~TA04<&YpmmZ8Jgpz)u5g3V(K>2R`@7 zFWMC!yu|+f%+vIF%qBHWwmi?&dcX^rOstjeRm^EdGH6)62;V@;AwaB0A$&@hMnmNjtEIF| z2*}8z>=CNjBwxq8-Ds20)}m$hv4x0V-!Pc5b(nm%;J+J*l-jta3C_fO5EJSCPJlRd z4Fc9{(cUOdsYnqM5N|D7vXtTxSTWnIMgut(Fx|3lovrWav0Kbo=^UFA;JQU*D z;ck1wK~+{4AkvILcLslDG{PA#;XhS-scxAtG)HBqgJKgarn&uBzqU`6PD~M_!thIdw{K3*J>}k zycn%EEZXE@ZLyF>&;xHg{!R9|Pv2lSf9uQmoA)_}>Yx7nCmWUc=95mduYU2Cjn3^4 zAHM(n@7q;ZUFF8iNZMNf#v|#y!oB-t4|otV0#m0U^mSWmZL`H$asSVqKd}HS{MG!x zRg9ltrD$S`S%H`TP$3n1;PvPMA0fiVop*tC;)}Na>3`X{Q{HSdu6Xb3QI);snGFb+ zl{nh^5SlaCEBj&8j}$qz4w{}KE5?R8DU(AOw6X}B6^UM}jj1uB%soY6h0Udtj!JF-X+?etj2^x+dqY`m|<0+?6Jy2F>mraDbz41c?ubL7&IrLwLLjwyk< zvq+_xNwD7mWSwW+Rz$nV3xS}&kC30;_(_b%pSlw3xl%ku;7(u>PIF{$KcPnf)-uVK zDR|n)k{5kb#-ifX$rG%zub;4TJ#Nu%uXsQ9;7Y#x{7db$lTWu_{`P-u+47~fYRyVW z7R#_6nq+U7ccLA2*wK4%?RQPpL`wJH|NYlat97^65`7@Rv683!lFraYInOe$} zx~8(uBIPASO2KqG9Jf&Ka;s`t3Dfd2SZWa*5t`+DD1lR&D4%@zhStT9fss5piP$OO zQ%62Yr3lbD;S?CH11<<0%VN{9+)6e@3QvJ9L&2Dh3ijEN^^_BOu_pJ&lIrg(jp8n2?GUCP6BolgpGlE$C+@x zaW+~fLP-&_9#Bs^^@nx}Kuc3qM&k~g7_ub`lk_n`1!=Z4Yx`SLeBm2zxM7=`-PF{y zP5HKzN#@c5S|^`;vMF@aIxGg{<17a1-~8q`&evIjn zim@Nyq858(J;6`E1y~3Qfxv@oVdmr@e$0~rFh#|sj2ne`TV;!vE_43mvchO(oH}61 zm=1^VMMj2DtZ`|IMW&>2)ZN(w0L|KHKy+n!jV)a80%UnT&YEKa1be0HdZ=T}d^r3k zsZZJ?=6p=3X_uxMC0UsGI43wPpzq;=mhW8Cq(yetEKEl9Q7*LAdJD#StT^6n_2m<- zJ6>WPzELn}_hSC-g|!+|Kjy1`OscDJJCo_O+%M9Yo|j>P!;&F+1_1Lz706Yi@OKA$ zKtzl0yZnQjz57pp{ew*y-()kV?LQGGWIio-k)6ShPyoi8eXR;73q7!nJ-`oP^W%?h zF|)uju@5%7E<&8`Nh$7+_4I7>Ers&8-c|cd|6r5m(V~ks1THPqco9IH#5Sx8ie5_Ktp=Gjke(bStqegpfpO@WzQKi)*e{0+`2I-RSaY< z?RQBg&G{r#RB7#%1X`8k8!jzxl=**nXeW1ZJO2n&DQ@zKC!V;$Us>kaKl|Cw?4pY< za`CU_TP#0grJR}kc0Ob;%PtQFAMbF0l^Mbn6Y4C&`!3C_#w{14rHqeGf`K!h2U_f$ zlcw79KVQV6+0gr}`vHV-PO~wHxltg~Psk_MtE@}}TafuFK5+_{hQ96|n>2S0ps|{) zLEIeDR$g0WqsESMe%-x_QI4yY`d+b#*SB^!YYn{vnwR7ku6<``r*%dQQ#o5l31x?p-P7faz^3tOjHr>__v9N;+MPJ3|-^G*d^3Sna8*Fs}!aE_Uv-fWAA& zM+c5IAT?8yF#gi zu3o*`{{HvB+q>TNF4q$k6&1E%!2|&0=#HwSr`k;`Uh6~0hpCv zz4S>jgFlXmcbK`UXg@IzdB-{DyYK8P-?-5}{+X-oXc(ez{o41LcX`paAzk+Vi|yql z+ioeSBDJVH7#(ymj5Wtxf8FP|`FYz=r9C!_OsKWccj%#q?vy$ffbmZ0#zG zB{u!i%iY%b!UEIQvt+RitXXZ*wzkc7i_Bd;FvMCt?Ye0>*-b z(muH(pLKqjA_A!sl47L#e)NAEJutmf;4WRW^;AGiu0(5I}Q=EP5I9tKo_Q?FW4Wbz~ zSf%}hbP1*UZ9-iDKTS+C(Hco07Z{fBv5)quPr%@^NQVpvaLgj&L5-F6{QC72O@FyS z$@$9f`}@4luQawRSFUsomd~$Do4@$QFFKZLkzh$vET7+xeB>i`(@i&xsAK-~pZ}OZ zutG%n@K+V_mP~U`J@r((<(6CQh$D`$mtK0w?!NnO`~B~KKcZ*t4L9qxmd}3?ya3~? zSPWvU`MX&RVsvUu%KumkURha##L&8S_1g*4tF09-ffi5L`W1(qj6Sro9u|mtYvCUN z@YX|^HFL^%tV{@*#P}*g9i2^u(X0}C96;Sn=setvF{$=A!aJ>}2hb`P6_;2~>pEMx zY7Ls}h~oxW-`PU!WbTJ!KmcAm1Ob_3i_MO4TOJw3yhI#f#`EYBwBK#(tt9FPltTJP z-O~I4ngWzr0HB=G?4Uaj&Sx$fnupXz@@gS2pN8RLXA-URxjgL2=3&K4xij9iR@NP} z*xF*tlufW?aXI-2$A)8t2S)t?Os{1P(Z^gF#dSmA+7*z0&_2A^@4-1|yv>d~dY+wp z)*I}h#~!razMgFcs$<8Dv%f$1w;k>wS%8SG`l9pRy~AbfEjQkJ>#cU%ZMW^T>K1_U zPV34-LHm0TRLz=g+VV-lV6v|lM%G+dmi+l3=2F{}roa0#oBYlTw)vJq`CIR*{jGnt z-W!Fh!+T&5=G1fyp)(S;GFY60A%G<0R(`Y`)quX@=9Shs&~6d5A1GO!a4F%~nMku2 zNX`-141r@w-z1&OB}tVWAvER4s1Pmz#2L&E71}8U4R#ji#5mV8;fN)|wPrPQEYLL1 z620YUeIWn~RhqwiJSNS=G=+rc5QtXpy}#87t0L)9`9Vx6{qp@KEYk~%7g|ZVglaTf zXo^cA)GFf{6UR3|ywzjno^H%`DGO;i9bE_M5ZY%Lw=;tZMMZ-!!}S1Mi@BcQcus}S z#+^9P5$U|RtlgTsx`}=4u{q;LnYSNJ(ojF3c6a?D0a3o25~#r!4m5ka2W)IZi8Ypb zA>|Q+@6ZauP%rK4yEYb0CAhxjEpM?u|M}0hY}v9``TWYy^x%UJc3pAYamS6gO9Za< z+x&sMU{vSvAsAH>wSz~SJ<D^X3=jDnq1G}{ z5e8Int(36=IA$qm05ZE&7Q!tYGin=3vEo=~FD+Wcb%^%xDVDi(n7N>e*uf7x`UJka zXs3t#jmrta<|EE>5_b9^yaMN)cNK(zVe4M{kRt#cgbTsYL}#)1P*^5r5;o=q7O6z` zrR2t_W+ldCZidoHcgX?ZJTP|Ghh7C~iumU zIdfT@Wyyvw?z^{t+g@J$lAV6)8Ft)J^B`4x#nM9$5SPI9H+<4gCtTCKV~*dhXYP%e zfBfSg+r)_z?SvCf*lE=-0OOt3m4$)|J+N^PXc7}S*GN+UAgygzr9$~z@7cY*f41Jb z!qwqDFnRK1i=ib=KoTRLz9eHfjX9ts%GkgyJFjRQn4zjBo*7X zZKGn_cEvlkZQHiZ-tQTwztR04?)~hw*1YDNijew>a|gtM`Ak2Y*oFxOg>Y!*U?a_) z$*tRDEU8i*y{HuL#9x2&9Ez5uRAvOqa1abB0l=7b6|HZ3&}VE=w2~p|Of$P5Kq_EJ zZX64-{_&?{Dj@s*^|^GNd7ydP0T*>gcvkR_2bXm}A`;pU1U&XCyHmH;Qq}4cUa~4;rUUoViXKXYfhLojfE0udG z%$9mhWwP;{h$%zlc|yW_T9j!0_3qS2h$E0HePMI5sL#Arn}&pnuZ8OZ3?x)g&eiiTH34%h*4iB@%3LsP3W!H}ZLKxy{xtw8rO)NP1zOZ;h=H$)w#m0D%US+3Q_Zy~GKJB$;`Ah*bzJ z8gm1}m7jiPlzBi#@!~fjrJzOQ@@FAZlAN0HVSW_u`xP(_)JUTz>uuK#Kdea|$>wsX zl2?a)D(q~iLkWIW9RkgYdMxb(b)wXuPd+*2Qb12@(Od41lPN*%=3*M3Lj-}xih}E^ z=ry3^7H#tryw>0ErR{lrlGD3?%i+mGi(ct=r2A#@yY9T^a|84Kf)2}AflU!HMCQE! zB#QGgzVBEV3H?9e9H^5sc@eKB*YbE<_Z!qO*jY3t+pu_pO^Pk=u(g;1FhB_@s6FG@ z4y~xzB-0tl$l4D4_iJ;;^IDVjv(Z@DV8WQCc_p+QEm3YIQOx7z-^WP0Ez@k`!>?KW z7H5%{h4lZLic*e(iGE5*GZ;Le^hSF_@c%jG2d|YvD!OH%uX_^<^Su#?LPv3ht(QjG z{3AuSw9z=p@B&E|r+V_-Z9m}rFi;_y33I!&0x;5fA-4!e;@SZj7R$OuW%*-g zm(coPUqI%QLi1^_x3_CTiy&`hMYpXpQ{A!`;X{axR8cMIa&gaYdjSs3I>= zrCUEXrqE>l1D)FT*Aspn@sPAD>)NON!BU{cos|w;%?iB8dq}Fzfp|{=l!yZdstP z{~RXts3l-?D|m&dTYbjk`HIP^x(<`|0B z(mM_t>>A%wEhF53mX-X=L*-~jm&5>n>`0q$%4arcQH}ax{wo#TPH8t@R`ApLoaC%W z@B2-y8HO}>yT<3z3h(vfYZu9HGotwc?@JHT9*^^`*If z@q^)rX*IXYR9%4LH|Cj>P$FU>Y;yH87Pi@}z3I9U$ag0l$fH$_9J|V`U@Ll7xOjoT zdQP>=AGdIotZz7Gg<<2;^oI1v9l&vje9Jz}gtkUHgJwKE+G>K$_2&?$Ebe;iZ6!*L z3tP0C19aKHFZ3WkffgupdmKLKaAmrjK?=;8kUVXdhxv=N5$1fmBsetY>)k> zsEb7QS0sTCxnbtcP70chr5N%-O2QoUKna0$hAM_7=5m9V;^-KRZFP&mu|vGw6;OomicqZr4qSj>Chwm>t*%^YI>&6(! z5;@*Wb}TyX>xvJ}=aV*ij3#5-PKM`e29Gi9G&a~-<)D? zGrMzH)_RG1)(X8tOd_o+Be?d$|0!JvMyhI3aV&t-<0gxCsDih8+NwnGA9 z!PyF#Tn1TkU#->krkCXpTngj^)<4f{xyWk%xy?^ z8tYG5t^mz3TM&j!MaT%40iNb{J+W^s@v>DZ3~bW4#PHLU`BZSo3z@{E=zH<+f5Rbj z5k@`r(jxea6(&t0tUR11GB-1l(o5p&^+gT7uU)*fvR$T;Ap&mN%xv-+amsT!AYN>x zX+ST`>-ppjsER-1cMyA1*cTH{iis0JB9mJV^r@KJO}Tp2UDVe9#r7~hOGZ)%F)T3k zX%9xA%r}6KU9~mHjODhA*8zg;K@%jU|AlWfOqaa9y+uzHG#BU3y-Dp0+HLXi?nkj% zpDp~G5_H$PnmmpVxsvmHON5_&C?xtEVQkp}!y?njm*GfF(kod2Mxk}fAS}ImmKzyZ zBAD)P#V~Gp>1ycu2qV{*^&6t^qw@piW6l9VqZerOV8|)aEGI*x@ZgTwGXUBvQ^rX%%LlP=KvhV5Shck; zNVh-$KmvyjSDz$aH0Us0l6}OtqMW)uBOz?Q$174wr-e4vh3TW-c@+$C+hTr@8_7;WBN~m9qtArm09K-*{J2DT>+#d70Kc+m?cbzTB^6=TgX}|uR-5DBtL^Y zabDGRt>ei?u`KGuidocW#zuS00Gj#O9=UWdz}$r-vg=X$g594AintC?^m(xWKN~Oz zO_aEl4W<=h(2U9K;|!r15GX#esSRyOTBM9Jbf1(XoW1XOSjwRjO)~o)J%Dkfl0yYU zFS;{uvA6DcowOrVIR7ttX;obCZI*ymq$Cwf&l@0<&N`^;z6)9UYfJOjofPAcpidn3 z*1KbQ@MCM*ZSjb?CyB?;OBb_=Nhh|*4|#(Rb!8Qexe39 zBi-S(8CkCt?GcLfI2diI_%$WRZF4auCyE0PE->Of2;YHuU-ywZ6uuNb{Bc7 zAnqm&drUz3{@PAlnN6nyA=QZr=r(M*IdnEbi}*=BuuMMSIoDA!tQFmNl(1nUBex;8 z1P^B4HDYnHs-ZFcx<}_mj*jBUp*Hl18H)=G)+Gq>nd2qK_ncSN< zUsdxGi9|~7yFML-#Q5j6>N1u5iT~{_)DJzjR{_2a1ipP`BgWj}d0d!ERUln0r&qF6 z-0%pJFcV-8N#y&I!VpqCyXr()8x?Xe$&66tXi@QaF7Pn(mo%l&#j^Sf2vSg`C@~Z; zWVWeveDO45*uev#q4{38V0jz#aif!SHwY0?~}|F&@Z z6{KoBRwGce&BI~0g$F4P%70J;*%gqLNK-X`l|UQa9*h-?5tq4|wGluMz>C=nb_S8& z{%>>$n@X^c9xzz3tlhp4Ft7?zvEU)Lop}6fp=GsE8sc4JK`+cEn zfIQuGrE)&l0FL)w@<#gB9=P%-%fhbkyC+oGF!K53~&2L!?`ZYf}M^^stxr$ z-RMand;j?;&%UC;<%bL=t|dmnW0|MxL95FzXriWwx>56_afTi+brl)zB~tT~WfBcV^$M=49PUx4fvOJ`13tE~l&sdPUlmT~O+$X_-cc|v#uj7@BY6{F#Jz1Ty z!n0`*$a5ZsW-BvZi&eX^QOy>jd%aODSf$1(BMLidei7m zAaA#kPfJNsrm?;m!d)=HTk}&NN?jAb=?f922KfWx0=<3V70ro0eFb&^)i*TfBxoz& z2sS}h_lPh;MBjv|6}h)1)?VT;xv+{#1$0FeIumLHETT{~dh{MQNS&R7Pu`znPz2I; zFU?*oydFQ6B6}KechY*k+^7M9QH4-LXRtu1g8>L=5>-XflPbfM7Dx^^D+OKF$f1zW zru4(Kk6_(^?MIE#@mc(HC+RvJu)6_(Fg=|0c%_7s<`v3@M$<0kX{~EvT-63y12yz1 z(WbA2eBBfd^uiYPTojv7bW_u;GY+SrU%rxz`0gh3*R%K(5` z#`X(VtZDa9!n#b`HJt5T^_63AfWYuXrJE+g7;d{~XuZE7)3r6!x@G6Eu0KGebu4Uw zUWmR{DopnX?Cy>`cspR~zVQ%qj9T1CH#fX;C~sGHbDCSj;$TccLlb1bgD`5k4ha>U ziE`dSScG7KYY=yur!h!5Cgf%Vis1BPT*Y~`98n*=@)v41+2i3 zc|7Q#XIB3u-xj#%k5?9_Az5!`nTz#3993mMpkZ7R?KX=;$?phqSh@xvzz%J0L!9dD z59I-KSuv%Nt}yNy0)f()SySs`pb!NGzOl^FtmUBc_or;G>EE=M5N4fijjyKG(ylu- zmbP5?Jy@@Et?jq5Q5}2qUNz5;7RpkVrGj3o;@_;chKW~Kr@mE{>n@w2CNt|(-l6MF z=R=)*PC04zW&^IEV3$8>Zo@XU>-Rm^S;jKtV;UHqQu zFIkBQjjvDLrr<@nR@*~C1G|lIwo_g~(ot5;etjsy33S%?h-!s%V=C!X`gu_SX_J^sHM$_xY`qmQ%v^xf&rJWowZ|5{d zdSaUIeuBVKwlVNVxU8}W67N*(CwrsCuDD%MAFjtwE^->DUIz5jz$sYbHlqOJ7|@Ow zEq{?(Ru4y`Ap@#tf@qQ}r4f$QT&FfFiB->%>T$Ba@xOGIP}tHP2q|1WZ4xJW%?tQn z@Nb7^(VrhgFjWsiCoK6M3+S&F4-%mMm4Y;ZT7(VeR^8){%$$r{Yk4Rl{3c4F!P(10 zDmi0deZxM_{U*lMF(hLv8H6WZ@)=}tdp>vNP)owyqXh8bqlcMpI4Q_Dz{8EzaEti_ zKeg92%ze<)H@P(w&O&(azh?~y8>*Tug#)VCUndH%Yxn$E(7vOnb7j;ft{M~l3LOhqjZ z>Vt-`6{1ka(6Z?!wN_33!Wtrq1SS41%na|Njy&S|JDT6)0M~ttUb5++{%_x$XwKZ? z;FaoDIimpmXe``)?ybG8a;2?mVqHJ{iV#8>LV-k6Q5J2;$1qQ26Vlv8;!&h!dRZkJ zrFPM*^cAyU(ahDdIjNtK>1h4xW2R9%2Y*JD&>}TEB>Q}BAUs*K&E_=Nr0$=1`;|OD z*9@DRS_0+M%YG~;sAwk@(d&pfKFeaY_gAu-ZmYsA>K=u$RAvW-d?g-4RPxqM2xrUZ zpRWyIelVly-Jh`)r7Hq{&ju}euJg+JjW3{zlXqXZmGkDYr+}p^aa~@YfoWgl4{>cM z1kdkj_fr=n^0)C=z*7Egz88cq8;DMgYL`%hhBjH!1{YS(;);GQ(X(er1K{oA2*kf0 zDj;c=@K3!rs~4uQn9&7zx@VnvtOi6RD*L7yq5?KxA=lF(+@rsix8_{DUC)GK5p0NX zNYGgKD`=j>>Lxr`GebuZ*SPaqc3i4Uy(z-|h`)t3k4%=iG+EjaLQnLCJxE25!Pp|M zTmm;yITU5AmoA*@tU-!3S9Q+J)+NI++AH9I825su@XQAGHtlEZRI7;ia3fk?8Be6*|JM`;qFM{5EBSkKJxMo0oI@LZ{BG>_6+w|GC!upTG0% z6Htp~^gS_&$_}b*=LGPP|MnN%`hdn-xn<%357PzKaC@Sn&bCJ}tb;E#Gk-WMcVLPR_isuyDZMG=~q*2UHFT zYQ~!DWh?c%w;B!;g;`wyae$wZP)ysmkIQ=3wkY+Fe-iv#Mw*k*hfd21W9nehD7oe> z(&eo1*i54c+;;*v{JuWo`7E0^-$&(q=YWn08G()eqUAQ-sPCNvYHcJ)M)IxT;}cg&49FyQuZqBQ)-308B?@ zS1e1E8|U-t1{d!%{cy97p2ywA~{xr)$SHV&(+5 zanO}UGiDA`r8AhNIPMZq_JKfceo$TCYOtUv^T$o5X@2w|fE=7<6ih84WRQA3O_@#pvsG~Y-EN@3?eh&WQEf$*9VXzPHl0kn#FXE58RWfi z^ZRi5O7SsdN=0M1@&S&gTZfl4=231QSXhMrcQggcG(eYk(a*2fX}_-B`2S=Inal|$ z7a4sIr+@qTp_L5+hK-5Cnhagpt4)Z!C0%Q6y0!q~J~P?(Nair>)cFEZ959()6~q&6 zFud#p_p;8r{C}fX5|B{E$?*d?2PMY$l94sO106tbKpqvs0i?>1)eLp7GW-cfc#M!E zW@9y5=t2G+w{TTjOq1U3dMK}lKjA z$)YMsY|L(&800^qq&difEw^2pJpIok{-4lU)A?6^p|3gG~$% zT{|;Sk|}DlpS>#yqxJJ|tV3C`wV35Fw?(K$qN*gc%b>uSzd(LkVK)k8mB?;}@*Ew> zU^ALNYuP7C?OB_(G!2c_f4;%3r`%v%uc2v~kEEW%5F0C1Hh!>4>TXj8q@;5Ib!krs zMwJCi4`;DyI&=~KS|y^DEaj?{Pn$*k1Ktd@VnF)ouRUAYSoR~#WfGi@&+b{-TLSyA z1zt{wpRzB7#X9;EZFJ%QNc2e|zf%m)1Z7k;z!&q?lD2soVHKl|z>mG#4>+8z*$)G2 z0^H^){VT1=Ncs|!hnvWQo!^${(o3`1KIRY!&^n`$ZF!x~LoyM2Bed85Bn&mS;10yrR-hTB zHaErn@5B7xW6pynk2{>d`e==By$}J}oGez7@*|%^{=JW;X9#^XRcR5FSJSXHANirnY>eY?4$x^0ZH?S!=b9Y{B*C#}(w=yqlBX1E@6 zZO=WUkZx!a63eAA@2bdnya+Y)sVfG76r}PUXZdF+k;n0~Ov~Q>CL02~j zgm!=sjQoWWnJ!!MC|DeLT1yejH)5?ehVHvy%E$3@m9@zPA3yz4egh&ib!f)kS@|(t zYZC&&@l=PSZ0TTWTv6_vdBpsSogj-(iAKM1MH>6$8_#+r&(5Isy4fO{T;nX;^avI} z>IqKHby>&FrpcNi0*=@!-r4=_>!)DzJfm0tjS?OKQ-9Two}3_HmfMHxW2Vr0FDU%V zI|6cidh5mFeU*umP)|+!L-+k)xSEKC%l6v5UgU^CKlkg!>~d^eS7TwlK@r!_j%VNf z{feP^HR_)?Ql^C8$F1O{kDj0(n;M4zq#kim0AB9;mYk;9Moy#HprW8Lb^_+Vj19^Z zCcUmrmfhy&YGBBK#-IPM)}*rqE(e4lLd+1ba@+<1AU!Gmw?5CCW@|D0PwHmXw0V^W z*Bo@BUW)jCp(<(y4XxH@y~r5?!y82${H4tptN0yj*$OdN@WA?kEhJL5(|{hTW$g>y zB5}|^I*E!j5gv_?q&d|(ruk3DlUn?nhIN+{O$v;{yddgYQ`20~arLaxK|yrs7b3h1 z&`vh7kf1$@#}jpqL5g|Ivu*-O+vAz<^`{hX)iILXnU{sSNy-P1p+f9#5GShOzxJ5+ z(p3?EJ=k^}XRvKQNRZkP4wW9f59iFp;(I=`@;=LU4tM&X;$hxN>m(~W(pj1T>27;g z+fgH4{sirLgNOd#ykNT18PqW;KKz^C5<~xR~Mk_Axt%21Y<>YU1 zYK>zN^@H)AFo@Zmak%_9Aa%Lg(i9O!OJmWPvWjaTg*_9I<)MrPH0@_rqCJu^c5Q^o zc_^#qbQpbuP}E?Z!O=4NtNhMsq14;t^2wke!&SN_G7e^aGR8wFvm}hH^mfaT7Ml6vD^d} zr$4Pu`?@y!8}cjfcm4mr{~gc=-|bWs>GhIM#FKz9t+AU+G@XVpb1;~cu0I>2U8o2Y zUSnB`MTr(#V(I~~tjH-1&-LHpZfVAR$QgB^AnL<{JHPiF>&4;MFm6cea*_ zF{vdr>61w9N`r4c6LA{Zkxxl%9Kbl#ETR&_eMem1=wYd4#ojAR$^&m0(hQ7LFR7Fx z#VO+#^lNQ?vcz#P|b6+Q)!Hsg8&B+iDP*c)H@4!8m-;H@FGMWFsSpcO@ma#%C zV0G^kkWUg802;03B%RBG_6N69DMq1ZDBbeYFu6Joi$Zdp?Fa`7JzLa!voOF&$?zm@-G;wqdZg_4U-u%W&7> zaZg`j(1Shu@&Mb?ma0nu({oo13@;TC?Kr0P1||1&18wbC#}&yybEfa9ASrg60o7(b z%TW4t4l;ATfzG9x){J1T?zJoagR9gfr7h!HjB;LIl4DI8ss_`@e(MKw+(#w5=!~vh zF8h+_va`CER^;RJz8`+M@fazhBdHL|FZN|zFW;D-pymtzzPU;RJ)}c?;dlJ9uHLY0 z@!=WP2f5R3>nFrBmf62d0;|64FXj0|?cU1Z-)ejdstG~(DY)*PhWlc_T{Z3<;R`9A zSj-+fKcXJ1bnPS>=G-Rz=B~6qTqHew=E0Hp^w;i|L~2wrC1et-EIxiiM+pW$J(rIQ z*3L5cbzDE=1sJAt)5(1cb?bbFBduJzZJ1)^Jlk4Xn{S+SUA{8=UT{VJmXrq3=o-|L z{4Zy)v(`R9=V8jtj%J+DM>lfGF%cQOQFt{xaTlCurKQ`=1RQ92P>g((uc2V7Nc}h-1@Z~szi1FF z7%URi9+GtQ48}>(+?67#e=wMa3Ix&hO#Qe%NY(>8cY`~oY)mo2f_@JF172aD{Dw^y z)VrR=JGQX49kq!j{iHT|JZIP~HjaA5AY_I^OekX&R4~IVNDssT1f57+{we>5<8G3B zHQJ5}Rghkzcyfw-CR_GYqfA}HSxtwfnyrYvDz6Jt2359Gk9sua=v`v>QYar|T@|(F z87RatIbx^-E3bTSPPuwHHVL7PqOo%}B*+uc5;Kuf?XV*TBHm*!Cs|chCF}Ixkb(LmFcpjpyOraq*=8anACOh~Zr$3p+U&npnY`@Lk< zc=pUOe0n1a-$55PE`=;0Sv!V~G-^CLgRd`2JbIcaQCM(3Cc}rl{la9|1vL=}(OhDy zdE{k#G=xC__rV@Pu=kOFzT}o5mhL$Z`d7MY@f5$yFaZkVws`oYZ~*G~yf=45_Pg!uy+zQqZA8+2ad$1l3O9oiR{7Hr&05H*); z>VefQxvN#pK9Sd>M#K0n8)D1PAxLneMMZIuaHUk$!Faar#ddDuc);iC_TadS^Z(wT z=LxZ|=nmL_h>|N^R^&460xfoJDCVwS>2N6 zJ_dvK6;mAjqX#71DbkUq*n;&S$RLB@{t}qGgy4o=Yju;8nsbdfkjPAnb8}6Atq%i) zPz4wOM5UApf-~dE0~Y?X?@r+>{Qe!w|CBc0a?}W*)zaKLoiDw%*WQU=uyfn+n)wCI z&9K0`{VMz1BA9aHh-g?|`4juOuYNxcWq(Wb_^SuL<~fU`|78#E^+EO>v}=wQ{(=Y1 zq9;%5cY?;x#L%!-cGaNFYAH+o z)6h7>*0$v*Lu6%;O1V1vr2tC^_=Fwqv@UH<4Wf#ZX>oaJf)+ILiJMDl^%>`)9=I>{ zY0`Y46Ec30Q(+SHdsEsuzyDnc0PXT+v{2`2ci+d1sQYCvGRB@!uV(VnM3c$S=#_kY z*xkzFc20CjWaOV)=YD~j$$M{h^J>P?;#Eglso+z973Y}KOjax9{P`hWs*SU+xIo7@ zKT(~Vo<{(I&z!*vhn!>P(DRzhWYbMCmWKSH6nXY{liTXbkOS{s*zzMG!(O<$T=#n2 z(VQE(R7O)1$MIRdhl5(F&UcW`W}xuv|3>TD;in$^?!8To(RJJQNkZS3T?PR~A$8$d zdm=X2VpL&@V{TEz?l{uH_;^5Q90el2B`$Pc zXO^2ieN5iBd+_Fo`3@q}*fsDx%Qw zgs9+c`t;gncl(G1PU67u>Kx_$>ZDGhJcE)JsJaa+&J>ZIZEiOsn@RezEyF26D=@Ot z+RRETGb01vt9q}>7A}IbY8p5|jG-~7bk||mikj9JlXq|A+2FhsGmNPPABQb$;471>r)6db2_QYvwno~>sX4P4l1L%i=fAjI#n z&PYxBNzoieQ3YVGECVsdDG)7X#yyBrP5gs9?Kx}`Q{{~dOqnz+PI(NzpEm2@y{yYR z4ehCAn}@nN0=bv>wSBP91s_k!cbo&!5DZRHP|ZTa$;ytbg`xQ&TFEBEcaMHF^H1RX zTE_?l&S^KP=-;p-IgE`9Yb0GDQ;;g}v0Y%;Ft9F~^v2+Y z`?;K&T-Qzeo;Z20xSTe3e^zDtJgsVZ`|s_Hxrp(huV)x%(6gDcNF&Js!NaG{b-PV_ z$pLjw&QK}_FspoF-yt#?u+?w3=T6sIuj%GgOa|xm5|gg|i~9>;q-wH^+%-lnWYlA| z<;HMrT*61*pbOxPj>dStR3LeN=(-#V$cI&*tqQmP@0H)QJ@x=|O}Zupyre3N#^`I| zO~)@$O!rVPkkKh{3@1;D3}EA1?Zv^2=DHj+q&(~^B z3-NCzbEmhD*H|GuH+3E`yqf><<5OW|LjBOhJtYnPwlA++DKN-g4@CY8n*?PrIvKGT z_C42HclTczVB@s0QHqcJXHv+r_S^$CKjTFM}`RN+X#<76~#u?h& zqksU=|MDXZOoAxTX45_xi2Obrph@q$MtC*DKb~7?gjQznnp|`DjWxMuw>r7`b3DET zIiCIhavpqt$m9Zgsc9YN;zJ2+;0hZoG`E*5>C|d!M_*{Jqu4$6(BH>vzHN~NpYqtQ zecuG-xEaoqB$Q^&^{oH%EU*#Yw@OcyF!43~ zPepW9LHeWBVWEgGl`d_ETVmucaL|{v^t-PB&7Xq9a$Zkehx=GM#zppy%8`1{NS0Wv zo(#=8!0PabbtknV0cn*SmyC!fnue8jjXj&n7{$^AHp4>Q-f+{Za(gHp$tJFk6&BsB zm|OP9-vQOI_VpfX4VYW6E$+WKkHWR=(23AgHRt)KW#)~#GGSnueHc|fJ48LO{L5){ zQ369zN5u?Lr_$ukyymUk`f}4IYhF;tX-xy}KvC9c=+iQFEDwHzC_U3KA2uC(xi1U{ z=04$pEDzFRehv^63`4Qc-pr*Ia+|%rDFLBO+P%gezN@ZpIhI6|sk>&8OL0Qr^DtY@ z2dX-_k&;^|WF-LWJ!-9$QI#DJp*qOrGl-$arQd8kiV7ViB`JB4m{)HMFr@PfJ*B)So3;9kF{Mru%We+miOvBp6Us(w-$o~^y2>SbnHNm-WfHXmSd_SPN z^&H;{ffa#$@L6TPaOU8@kI#5}T7=*{a=ao6hVP-9hdR{3R+TX1#81Zv2+48Rg{MI&q1N%jS<0=%37yN6Qy)Z?-6 zj4k>vr}Lwlvo>WZhLY-AkAoY{;YjjZaAR!bPL)qPjt+TZ~S%vR0hPXMr6jIW;wN z0!5Z<7Xqq%IQ`VTZ*zP8+w7VjguJ5s<9l5vyZ_+;VIDjz5Y*WLl&r6Oydbb*JTLqg1K7v8Jse57dXbA zXA{cZt#PtMN3CBc~p8HOnkiPx6r?WQNe{1T*=>&=m~AHc?ucMKh)cDC98WH zO1^vXqjIpu8wb=yo8WyOB+uDN1K1bL*_sO&>kWEBiz$3oyK% zwCuXbJ|3AWJW5eFfEyUN4=TGZK27Ls*X|75b9+>s9%k`g^5lMJ1XsF`5x7m!<^DsH zp6I5Da#?FVHZOk75syFf^Z64@Nh{K_!;w4fqpPF$A_5qFVPN!~7Nvnk&im-Dt))d_ z4(*snFhPZ9UB~khWykGr;0Y)ypYg=hEA0@N=CHOG5h%t#U>d$PpFr)aJA8gtwV6?7 zFZC0+vx1q_)rpp3aTu4ByM(L0d$3T$`hZ*&Jx({AVOp$n321vxX zKv!e98Dx@DFRkOomq`iq8*g0%m|QqU87;Q{;d!)a-oX4)Jb|&gvA4sf4Wf@7YFz%H z;k(DF!d5g*?94EN9W=KL6xOkZw&#rXiemc#p5OAdBwwuSnqDW(HH6b{T4_4!5NLuH zuTiK&jNiGxmHM3qV%>G0ViPU%$VM<*HrwM7JbkO{uC*8`+n?>_?VXL*357O7nEbGD z9h~(JG)ynh=s8VYaPL}MLi^n%ME0@8`oUGE#4z)*vis_?sStV36b=i%i-|0Y=C%`i zKi)BgmJ)nM&1^M!r7`<;9S)r^2;2r?#SEk9iANuPf30El-Ny>DeC*eM!}f>CoorEo z8PP1@B1JQF0Z*`d-VgR$qllnJM!%+buRCsl1{W?(Lx=%I;{y*ow?fD*Yd;r#e8W${ z_dQ?+H0aVBgWpB0z}tyEP}}Cez*!=kf-d*1;$H^`R&!Y?+ZJ`u#B_@wa+4$c_yL(H zlSZJ>mNlJ`&e>RUJ>IkO43wd=}IiFA$O66 zLt4#Z9JBkD*o;}8ar5Qc!ciXh>hkyl}pRv z(a(H9u=+2H91dJYvML0e7`>>Y)esKd-HoN+7!!C{t3qVjVmB2~((NjWR#NEK?skk; zm!_b#B|Z;-i&OR{@5Du0q4KOG0dTyY>^NmC2s3H!k^)$qW^_scO~%TALJH>w_D~N0 z_8}+~Q5<4xztwITa

KyHJHJx?)y3px;XJ0a>cvyJcXaA_dN=dQ11n^m(0N;f}mP z|MTt9;LE!%0K8?sAcm@VZ9D0?eQ#Cf>G0#oeb7TvDJPmZo z@x!+1XD`gI2L0cIkl83U@+d2FvUX1?d{g%X9C5yRNJ16hM&T4h$oC@f&Pr+sWd>T5 zi&F?D(OGI8zF@Vib+7J^e~bf-MMqGd(-2=A-ohvAzuEI*N`03O@`G`1O<1^z8@~U$ zeea2REQ@$*;U$GBrSE5VrvUzYWhq+dM6ra)=G;=R`h=MN2FxE!>UcJR2E+AYkfRzD%ybY&PUlCcdp|VKvX|8NS#ga- zBv7OJb~Zv5V~AK2)ZDTE10~GwZ9f;@$#i-%S?WV3HeTQ|%Uu8g^GC8D{3?n52hh+U z7|4y%pc)6K3YTCddTO^f{{4@UH z-B`z`)$JMR00AOh!|T!`^^Lqy5lj&|5!I44J<+ivrvCWb%U|R3$cAxkSWPPD(V+|4p_WzSNNXd0ht}rJI@%2n zFTtE;Q*?D93W_vb`9NF|w%d>;dWMj{63&YuNPkSW73sj_7_l^G4x|G0s?~n2$LQxi z&Z1dswRT;aVn|k8Yu^=gg6V>GqS74P@;~pG*-j{`a^>uGIS}|cLLj)+7HKM#6|kF* zWa#X40FXe46m=f`o7kPY8zinRf+2*EN>3P41S%yNBPTXsY6NL#$t=5v}B zFHVQRfvW^FcXNk93&+0FNWz$ z+$dqFz_{8zSO{Ag`-gzX5+03utQSy>5*|NaDKK zaf)j73fRk*J<+2-=d^-JolI)FF@BQ|p3|<|@>e|egm`FEjBCdPvDr~1ZJA09K-m&X z^8^P6rfOwF5@dhEv;=B0=9qnH``)BQ4wN|&!3lhLm0SRPT?6g|^7%n!qtBBHw~70D zaC6gqeRE3d_*cL3UDo!qm*jsWb;vkVH=i}bkbC(u21M?C%zfV}lb_7XxSm}&Y-7&H z0M&4j_NkMUlfY$?+JKtpXWeH?vEbF*`?L3s4hp*0O@8d^+=(AM%M_OHi_&hAT36WG zD9FLIUYFGJixwd{-fEDi>BzU%sB6G&iA<&;0S%sc!e0pV_t(6?h;U0F^E^&KXt}+d zG1E)murFq7Ym0}vT9BZ91Kbo1WA8rm(uKiR`jMYwROW`pPj#i zA{c_5r)Yygyy2w=TBB~1glgX(BiEq)aHw)S)eTDn<%&L^1`XEXl6DUpd9N9bi_@p6 zk~wgNHGh5`IYA^D?wtyK0lo0CiQyDZs+Ocfup^w2u0NDG5n}iVVu;D1u z9ogg`{~&tL%g}`jQv9_x^7^r}M=)lO$d)ygT~Cbf^J3HoJMH19{uMbLI<)vn-3Ve< zJ$Xvn!!Y-UW)CH3=n3$`s^JV%$Wq$HN?Pk{EztEC`;p=jO!cz<^vsy_g4%}kHGmLL zt8Z1qi;A6&zX*L`3ZQpujaJR(F=Ec+To%flN)O7vXRHLj52ysg0TYSYQaf4Hg_PBv zMdNHaw^1~JYdGUS%j8K!WZW2K4LIa!yRFXFO>a2N=_aJqtfqwlBcOZtlYR6Hv&2Kt zy}8!}PGI8DGT;GC2>luNNk0&-Uv!~DVDCTh11OJoZTKXNyF4A`MgWvej@1t411x>_ z`X5&-$o{FlW`UJ;KMb16nNu=wGAq>Lz4DwzxXw6!-7ZotQyhu-3p6P;hpX}&o99QQW2viI zFIq>!bh;cjxv-(asztEv&1J1R^HGj-6gQ!IU_I2AubUpn63zbu*gz-0zJJwy;ne{1 zGhT64ECDs~B1au@bl7j{{^6=?t_(Na_=oU!hd+7ZE|2KXd&;P>GdI0mU~F3dgc6+8 zgiL#ftv97t#_WpzFecusooU2$+;`#+{lw4H@t*7Vv}bzYs;jQrf_>22+q=zl^_H&q z=tn>L5Rbr~Y2735tWxYO!D80c$>fut{A4upIghDy)@Azi!2bL1Kc$-)Qz1sF(_^B3 zsw)Wnn36{*r6{3Nl=YkzP#j~OQ`Fk1*zzYTlpLbs7hmXxoThdjPNoN_xdZR5URHKi zxh$gT8f#nO#6^LkrJ3-m?VcJ5Lo>|>9!r9PXUx??g zt1%y%#@2?RALT-RG)v4fiaY2uiw0ih{6!ShA{^$#I~&VI(Jqx_8*eUX~L zoe0#u)}R0U=c!(Dj~{P#`Kd@#As?T=CX6@nS!7%ZF!hH*U=G=?=1`!-*o;+!p#x8H zjzYg?T5sX`7WWBmn;B;&g3csvP#^(0IUh8j+AlBHN_AK zW`joh#SHiMNd)+;V_J;!ng%n+yG!pk-oE|gzWg<_b;fywf7Wm>u!H;W30b;y>7F?4 zE}wHOhk=uO-<_G+N1+*;r|AE_^<6OpWwOR;Q0wlj)kMGLFE{UXz<9EIdbrnZ7Z@7^ zoru8Lc&)E|ACuAwaW`JzNVr*v6Cp=-xRN^LinCNjf z=-U1N_{Tq{PZvukcrC^F!E~Z4e`NKRX&wFF|NY;1O=C~e$*%M$*q$42yfH{GvWB3E zCnk!;rKO`v!xa1*8@#(>kc=<#`e|9de0j7ETd&Y->Ww#cJNW^(e}4Y+pNBq@CisJJ z_hmsE|6hrRTb2rz4IyBF;mON#_;CAG&nOfclt7dTZ}r zj-dn!cryT1Sidoua;5_nPw+T~0*eCNKzSvK4dGn@6h<#dEqRTeQuJRu&ZvPZjHreJ zQ-`j#gd3qcQ&9y36G zSyxCwyAc~|{ocgTPL_kYK@J?A?6U2n8ra~XeT;ZBk^dMy(cO?Q#8U_YogIyQXy-se)j zub`%xxxS2%DrfL552<4#mL%rcce?<$8acw7!7)4NZ52S)KqNzg_cBp9gljNwq~AMU zGzzig^3kyy87R&hd{(@so>swk3++8%VRtzDgpM#9_WmJ_^d&I%f@rM``#upU5rJaC~Z6= znN(K&-&}KL_}IlC4uVoey7Rg3eBW*XsJ;95!VG8*s~%YOkb25VyS)Bwcl;%OY-?r0 zuFFpM05ib0#n_j$TzTb{F~Na}9L)Y_5(g&~sf!bpWR+hjohwjMO^V=;q_MZsj0JfK zU^P}*<=-ENM2r&X_YvIw<~P5I@1-tT z`aqy(8xrN*b?JCum1R9n0`KQO_qkCSf5$uC5tBQ-_r33p=bFkN4FUi5Z~qp7dvcyL z&Nw5|snUSiQVo9o+Sk4oq|zoQhjB=W2<{CBk*02@>a?gdx=sV#WtUwR>DMY!0(QrZQXdoP z9^3ohXHnu>e4+X28EqB-0tE3Dp$y`oL9@%qOGZ`HyA(56^9k4)0kNJ}6+jbOsaO_R zZ`@#W8)|X8Szt+|g|?~ED;2rGHS)ZQc%%yrtzqL(bNKz8cZcEi?Bh;QAg%KY!Y0X0 zB`oe;62A78OT*GxU14+>+9byk1sVG9z70i}q(SJwITYSI2TEc6rn^H=dsFD9xibL2 zdZ=sVmZAc5Fl$C=odqRV0KspJrxg4-R((mMcD~T~C53fg6O+UVj5XvXI?BX8uy35N(tvs5 zK0EqZy~#N(?PA%D+|hd0x5b4Pc` zQ_3l{04jZid4^?vXejfUEwe(v+JMn2!`O>s0mGCf)G|y2QnUraMPhk9HiF(b$n>Pv z&Ibl|XwVK`am+JTdEXqb!#O>Wc3GdiF<%H>=x)zG=BeRdq0_D(3gPPCUmrg5!4HJ> ztJkpuUT@~F3i%4wZzzSUetk;zf9jGdadz6AX#vLa0W9Ua|Ydb_?ifP6Ydc;@EcX%lXdrOkcE@Im-RmXacbqoj@|c| zrysjZ*V)$(@|oVI7i`rh8&2rixc^-8W>xiWe~ z1+j`=Ug?b!MX%7d5>6OHnC7E9shRjH0mYs#DLJ6`%>CU*z-acmM7)>ehe@Sc537L0 z9}$0){GsqT<2O~prz*>rFMr4n&lkV=#n?|4uul5Y_--@m+CE=;trM}dLOLett6%+U z1T%X3e)z*5MrB`sqUhY&B_}pk%Q3?3gx4sp=|NZX=y`nM_PCMt<2vH9GD7wScFI@wOzh|fvHqxe1^fX^t#?vTG z9%Y$DtMw%<+)UZxeJS2Jl;Y@xyt%mP+N8M}jQ8)kg8ELau>Go1lj7g_5WTf@0< z+jYMQm26*V?aqYbo_$RCpKpINJomU`p+hroG~{5I#AxP#IvB3|?W%CyHTQ?U^{se= z%g~ui0Ty;G6FwB|F-Ccwv95=A6z_1Qo9y`9!x;0Fb5#Mumh*sKZ|rMv8RrVt#SHj> z@g_W`%@`UQMe`in*v8E>UsmXk0-t5ZWC6{2ENU7H_eMGPng0-8aQsQJ>1Kfbf0z9r z+;G*+2$-j~;~FYG5h zw~1FvP%B{7ODMfM$)f8Insn);6wjBw^ri78dGA`IrK2aJ5^vw=A(TRz{-KpluIQoj z8fK`Y>rJm8`1~$q{vW7ywoO+(cq@R%utTR z=#`YN@}3;i>Hd*DF)>s^Pq`&aCW>=bGQQkq$u$`hiXj+!pVNT8J|<$>Dc#5($?}z{~dWz#8j^&qCPjP8pJIXN> zS)guItkD;%D3Hy>HIB}}4-Msb{a9PL<*wndu^$>EUZ+lChDXixg)&*@M9r-LQWW)n zd+Wc3zr*`e7+3?<5!!UN7(H83A2)1xKC$M*Gkw<0^$td z&OXr5)4rKrU2g3NnYnYq@Y;T!Qw9aky<>$W1aa^<&K>f4eVr?+7^Ny$o-*uX*7*Y7 z*XRXiv%`;Rf@hrr5N3|JWJ(xshuNuRT)4M}F(M(tigklwa9u7e18mRZ&=x{EM(()- z=E@L^5wlO1!tCz0@QDjQ9yXl)`f%-Ue^WbNFEzE!OSpaST~P=dHr*c{_xPFNpq@_V zr$RJtQC^&^+!(kR$6B;(ORM9!3-9g-o@Wp-mql2|SoJpj>^wkA+24EB!_B>T(-;vo znpL1=SDD)c%AUs@VL9e0NZj$gXWSd*_cvS@y|n-R*{_Apjx9qR&il~2$Gyo#iyhv+dIO^|9t8e>v#2ok*1JKrzDd^t4`^Bd-`C10=FwtpjA+%K^vzsQM5iDbIdUj zWJ*z1c~*%|M8K6u+FO%&Y2B9gR3`gwYkjiM#N$*}1?cJ8qyq5-q$fbqq*5ARDP>#wwN}Q7wDEZGPW+kJZVeAk)QZY{A_A0zI-6eab^&f8#y|TV zDM=qEF_FnWPWXyHu>L?Mk_JI(z>)@K4_Ch=ZK^M8J&-)NNq;1OH2J(UOTEKM*n!?; zpKB%k{YzIC7*BjYlX4h7qbgEKJZ?I{)l|+e(6tVw*G-|vVd%Fx7p9fK-`_JN|I%51QDqRp$I$6W7vBK0wm^u+X8j-br8wxZbqG7A~v z?o@p7cr?JHETcqr;fY-UEwKtR9D$OUBet)B1U4;tdT^H-wl~&_LJY8N>TU|zmN6Wr zD43i_g|W&x%1w|8N#rxyI3ql;VJzHy@21eNCkTzZ0fk(8W)<(U3bf(MM@LXh*_pJmV=%Xh#1ap}pDvNYwt4n~FtmOhrM55Ze*v<~3f^MJY+F}n$ShhIMpvx~`Rrf- z>%|P!8m+Jeu`dA#j0N5@_F<7QR2+9wC&~c90&Z6M5YnI%@iT`+an~GwT*lue<{!=+ z$5@uP>tK95u(4mVM(AKZX{O#$C+BP#>B9xvA1bZwk-<`>u|?*hk6!rUaQeTz0q_5x zna7CdhtpkReTF#cpCB(7C?hA%c0EsCV7_^ghUmO$KlJ46A`m8|O~@&TRXQ%$;VuE^#x)Nqc_GVeGW` z$;M3^!pi$sgpYsVqKAOuD}Q@MxZ*dziZ|uFcHsvv9tXv)JnoohhxWGiaM#^;Oc@w^ zzC{Zbg-b5_Qn>4$JHzdF{xyPQY2Cfc4hV-m?(kq%?A;}T+Pw|fPW9e-e0NM#p0zTb zdFGihu2UPy}M5ECM5EQ)NJ8tNx-mFQ@;}v zqbX?sody|6lC(!kzohSFUwNNSgyuc#`Ijy~lWebPk#4M^Dg9l#-5=ZD-rit#c^LzV zsPrU&IBj6r5wApAflEJwQ#rrSWcs{GJEbx^pO`Dam}WiN(zVmqN}lWFn5ircF))T$ z5&X-%Fpq)p`^I7?QXdmZPfV}nDx7AW-K}_O^D)M+1=~>x1yXQZLU}BYKvNtogzh5y z;H9ZTMa=a9Ala5BbMp|M<}6^cJR8a&0J#ybO)R^D!dQgnSjJmi!no8okPCO*RSl~) z)H2dm zj}D{Iu`2~iJwaPkiC3X5;eu*{`xxVkqf*LQSs%+Vsq=a*KeN|o&h-i$Ts`y)AxQsbiG(=7`w5wl8-`21`)$&kNZR^8v ze35uvMjVuBv!XNZaL?LD8l%*sFiimHazsHihk45n4D0%D#n@AwOQalDRW4N3|_v-aBjUTrTXzw=yZx zj*eh_DurP#F^rXMXlRU(1)Loi+K7M-c~~OtEi=Oh-hW;=`IJ-G7f{T_^l%G~xwTLn zZ3?&EkqaY(4dKzt=7gCo%s(9^?ujrFz;20@_Kw*C*J&lTI@TpFagEsa&7mTk$IuBq z#FnBH9haHws>~4$vspI-HVD^bilP)2bp5!$bZj#4Gk?So5w5Yf$qj$HK7!jPKH-U5 zG~L8y=Y8;;@pZ3$`B@K1j3p>u-g_X%%$jp*eCH{?Ip}~#1z(fNhc0-3IPKI|ZnLb^ zWXILlY3$jK=oOJ#Y_cdtuZc?PFMjch7_)0EZX!~b6_ewnMOscG0?V|Rcx!h8g_?BKZ@HzyfBqaA#`HH1^-5m z4C);~5T*nG#qvo-=)f?9iLnLXbydR%C~+9Y*f_%iN&LrDIQiKFD@S)y$=JGTXBZqK zG1dwUusD3iphBw&mvKE=&ZttQ>=KG>6BO)9)GTZC#m?5IaQNdN9r9%4s+RE*<4rD? zNuX11L-A}2InG%OT_L;hz))CD*_PaB=p++R7djYKSOCCyf0~KxE+A>j6>; zMx~AKaXu4M2q-`+Xv9TToXdIVqTFH_Fk(?%xHTebVh}xs^O#8o9XiGxR(e{)g2N8M zYt%+bm{Q2&LKbWY^0R=D23$rtlx5_joddnX2ZA|YKhO#IcEj6}lkTX2w zvdb=u>ka$TI4>j5-}0%7wcAZ%SeIH4lwm3;{$BbVYJ3npHUJVcjLCcsx-rI^cC2F^ zGkQY%^4>6f$Gv!h3vA~=5T9d4pr`RG#@9Q3OvHsjtrf)F#2Cqu^|_5pi6K3|VmQJ$ zK=xSHxt(bydZOJ&F`{Ay7i$2eSvW+^rwLw8xw2{2AN zNUNh%TN85$lodU|ekr}Zz43GMC+*RhX!%S^FD3gigA9%P@AhDZQx7atNZOubA|1#%j7+9 z578-^m3`regS(+M6T^%0C$_~I6FLEbIeufzyqW{kqIL4Wr2ZHBNxp1$X%q6q4ATHe z%Q= zV#g_md0!rg1i;R%O!hS_r{p(cnX30APgB}7AzQC<4t&|n}$g= zHdu)UH|LR7=Cy23AfjDw3lP}Mmr_uDtM9|^GqLr=+-bWmIIi3 zmwTTae9)tr#>(Lv-?=Qj>9zkJZ{UU-{}4WU@rTB5#zb2aV~^eV|NZKxV54Gq%27`n z-*;Ev7^iHKufN`L+peCXe%y9?KvU9oMIbp@yop{(>D7ryFLnYJP+9*r5tYV7(zH#L z^5vU>XF^#{_Dh55WZkYVEkTo(x4rFcG0rKyPQWgJPPfBM(Fuh*+1DSkBso{I?vb!O zP!wO&cVmfwXS+2mzOLN1TpoBL(5d*7SZ8DE=@m$&T` zGeM{CC;j}oJ{=36j5XJ5z%of&S~RYhsd`ssPn_SxS_vq&UFHzby1y++Z%=#Aw{=W< zn`O+>1_S3J*QKRCCepnzecorGG~ctf9M+N&woC@D0vc+}EKBg58C&~NQa2JAv<|xT zdX(XQ6wXoh8NoF>M538tKk zT|vWe4HE@QcTU8-4bDYV_4H~iPmGTOZ|Twk?qiQVcKqdUzWGdKA0_Qe+A$rM{sszq zH4a$sUSMp9jO&JYhKKuycnyb<(Jaq3iSeKsFF54j(7b3KV}o(Z*vEMc3j#(2?qah~ z`x;}L<;pfmiZu@sV&>L%*dcHni! zJ{1^VaomoOkv_NV(acNT#0C?$Pol5F81>M}Qk)y!7%KUGmg~Yyfa(*Ua4?23=M@ZE z0A91^OB*Ri#)&b%j9cm&uVoNl@u!h+#jS;~akw;y4i$aYu5(9BX8m<5G< z0hy~8H-e@N zd(alMS3On-K!tw#Q=Snn_|!+ki(dAE@PFR*mT=mcr$#TbH)7W67e4<%R#|d$H{SfG zSi;FH)y4)-RsQYmuX|WQF*#8;8*X-e={fc$*|vt6x|36gS(pqErb`AT8C+FN(;ntD zl{x)hWi#=@B$IY>$;9iWw7|#(QVvNeoaB7FvtN>P)|Wfp19}wF zF;|-&*T_RgGOs|OXE|FTb8PvBx6aQt>oP4&s%%HGv`m9!{q#1GlZFNCOxLD zD{S0vI&VPwx%PZ%P^-eDs9`&spaVnc&;vp^ zVtHtI$^oI{;8`IoB_?`7b7$)T7R<#wXQ zQ~d?)ALX6*1rqV_GXhv9W|_+wrP&W{a)1OrBW%zJ?szTGamH{eqk!Tu5X9O|{UpI8 zMjIo;&`_4&B0xIHgj@CrxOQ6ha45QZo^bE^673Q^t5qw1>7@o3F zwX>k$F_|ZfGwc2O&p6~JIMz5;V-hPV%550ju(t&;a2T809*WDChvL$M*Z?|h>Y;5- zF%$L}<7aT=dTJT3Bq!I#ux{0gaKzz<)j&39Hh(nK5_|2QKG1smm_)0FUgBiGbFOtR{4N?sK0T4>HD2Z;o>0 zkw=cx5B>c1x4%73^!MI-Z}{H#z8A{@sFYg*MX@YUx>SL*BB`KyE){E&s{Bj!P0PgC zKF^;Pw@-{&R;iU5nZBOqH^GEd8UbQL2X`M6{Pgt^1cpc>w+(_}sZ1)^X^@@1cKUsy zgeU7%d(Nw>OrPal0|Y&w0*s9^y6j zM^ZJ&SvFK3>UkrU()Tbzu)m5NU|fFcuFIR-=j4! zU;XM=k3UmBpMu1#iDOPGa^fi#44aABxMbs&%~YLuCMG_g$!F(*^fi)a&};6@`6rIs z(r~*?tvy z=Yh1CgezggY%RX8J|@yVHa*@qlHe-W&=T&Z!06B|<#708=7qUL#Why(f}*fWZ>-?? z2whDf)7~5!dm6%RE6d@|+o3u_B~-zd0!>XOI>co;DyOw2S~UMND5wq5djmz2U3~T$ z@G!;qV9|_kP^_DqT0%2km$?A&p~3s1Z{|Y@?-1_f(1C}zL}uD*xQlwo-C>Ymzs80E zKswaLG9K%er6{yyhAvR&n3~6fGuB3mb{mO?q;yJ~&Z2qe`0Xgq1QZZdI7WPQ(Gpdt z>Xd2acob&Zl8Mw(XxOx=0ZzFkY1R_ooM}R7)@6n=PmfTorWiUo-zbW*T`)%CB8uRI zD!_MyWMiYa*#_x=nX`*w&ioFN`{04*ACa8gG&jDFiH7z!zxmBExu3W0Z}z$8o*RFk z-If+Zq^2se!35!FoN>l@*}lon>2`Gi4eZld%osfYgTe8qZmQA7`)=fOk#a+gI7-Ga zN;UGFH}PDTU=C#q08#cafnKGHL~?-Z5}+`H5vkBZ7y}@>1d=qS*70AGZq1PGYsb38 zTw%_E60!E==Zf=!L!IHw2mS+ujgf$GQmb*!m?+Z*fpVC^hJ{N*=Y9uLN^TyElXm8x zPQn#ngX9KjLl$68qOi`MFak@X*w7K?lNmm;Q5er=nvMo$t}lh#_qI7E*P;PcjB<@K zS)gLCdSEoHT{jfw&1ncrmvn~tGcc&xa1#wOPqep_h>+hGTt z{uxhkCi5Vi+O%n7xbo^N;<|r%;VI*{d*@wu5NC{KVHai5g2gce<>)6rB|PWYf1LCz zKF?m`6)!zKbag^)s>@FIz=Pl9HV)ORD7{BIwY5d911fk@=@d8#j`hH%X`51YomlmD z{p@Ey3kMuvp^c3{lXy zqi4b7P>u_q2R#mkF6el=_LuyQF0cEH#YCG*A%`C<-2B80ZNVU`~9_ zefE8RRov1dGvVZuPagL$?8x~AvJ;Q_e5eT(65;Vfar#iK%&*q)moiegG)@0*}f)vW{oRWr*dvm z=W+gXUP&{AG+;aDO;t3iYW#Gm!9CFdlJol9sawgK+n!fBIc9HOO5wLenug{=As<1E zncGLmrf$sa%x*Hh^BY+OKsS+qu9@E-Lj-b+p)%Gbv&U;3YQn4BFa{W;Gt5+70dShY z3+#y@iLuCBF$AzACtVyb%@~#nIk_T+Ix%b@b2I~RTR_WY=9f~p%neE{EtnIEl-(%+ zq^r%Hp=s6<65(})PJ)tUcN7KTTq{337P5GYD_E{d10&F`+e6>zFvc|6%y|p6&*9|| z%>z{(Z576!-?#Hmg|-TMOBm1^$sJI|3(ZYXySW_h+b|MlwoxJu%Jlq&B;xAKgt;W- zGO1V-`;F1wDtQ%3Js9K&!N|4203kqGyPx&#s_3r`jQ_QirMqvPeMCKY+ORi^F~nz` z{<>}6f_01E_pT3YbDv$=s1Edk<4+9MQ=ZJApFXGcj%(j>dTssvj`RQ_Ob*|zi%O+H z^3X#MjXSBls@SSr>ZNi0#I*kt&*r{=`qQ7{YGY`&wZA}6U@B29NV5 z1PJVH{gS?)>GIxTE12_Hc6^!)%J1FRL@NAb_pBo3m~?3wB!Tztl-+%Pl}3$iY4B?fqxMmu0N6n>XzpK6yi>(<()I2X#=oZn`??Vmk;?7Y7|q$ zbZiJi0%VrL0CC%Rw%TS?!+rE&$%lIo+YX(^|mPG-hmcb@zlkfH+qm=9u{YGT^q11*xS`ut_{{nJ@z8 z37slTW4v)5C6$EcuJ(}T*R~Cm3 zhNufZ{X?kEj6<{SaSs)|*?u=s>jx>1zjvc~58}2hAwff&OYC|YPbi3ByS1$;8l<1J zXk+NaxTR68)W|i6w`f2QZDD?Fp{RSlJTt7`)Es_y;~JPDl)qx`I_KT*B3H<}cI$V! zJ5Ki5w|yg=aQd@$$Eo({In--je%7hs$Rmyp|Ne%_=S-7?z362phKoM@sjzIn-bvf@ z<1hU>y!K6txSu_ntbhysL(no?}(_NrxQQz z$T2&!j}{jx;X5L`a<4mb-aWOC-oA;T*lTG8o+x_mHBk5?Y=7JDzSnvxWjiT_lcsnl z=h)l&C299`TB@AsZSb`#BEa05Cv;cNH}RO=eSVY0OceWIYoOSDZNsjDVvpZe2`RUy zr9LLoy|MScH{-P44!y1wVqpU*WcxwALnO{xyDlFV&&-4+^O{43gj08}$%d7HMRH0* zS={WEiMP%PqHI<`H-B=G22~L}plj-um4&CHwHXex>!03J*9Fa)!Evl!ziRp{64gH zkwMnj!YT!rXeN1h(1*EKl=dosz76`eN`8)b%K>U451_!qzE&lsKa4xZD?~;v9m{&x z97w^Y>#56 z-y(*%Cc`lps2R$+704&h;uymB+s@c)0FaIhY$U1OI@)G@FXy7>lG*UH0RtAF8E#BG z7`b@w{1eA*M8dA!3oukv`K^SL!)f3Dxs){JKy8aJNecs8sEW#*!6 zsWYrtJrw?W4|5Qg@C-F($c|huUatOULJ!=t;+}Bk8(&Gx`*8T(AFd5Q{@D*F>{0uU z24ER&>f7{?ZS@s9)&tuG#yfVJ`d)ie4;aI2at|jtW4^7GI8mg>e*4?sjwPTbZmWN; z_dvY|>OC;6djQXfxKC}filS`R*g=%j(G6v2%qX$c_%usxg#bW3PQGfQ`d3pTS&5dC ztlzvjwP?`0#AcUD3trovnVsR{i!TVze9955$8%Kdj~Qua0B8lTV^ES+ud}&5e&fxW_6j0vrg2=8I#huz{*`c8#39Be=L=vs0Mkv5-j4_^j zWA21sKMr%&I8(; z%CJ8OLAdF!_XJqqurvVkunq=bXKJ`AG50Ka)L~)Xf*B!uKY*S~74hZ*#~qdYn=mh7 z#M+wF5@RNF8cm>$ii!B{S*zaTNJ&R^U)!qo<|9x*`gywCM%_phq@MC^e3-f)}eq zEMYSUHirv%!$z`R6CmCor59Z~k7uZaq8s}}Wf*X*M_rZL_L@N#EDfy^fZkH` zf;H1GZLDZ*Z3gVJFTkRaMs>8egc&`fVW0=EH|cA$q^1?<<0uP*YbfVLoc45=-rn9A zPn_K6f&~lWb7fVG$wU^4Iad^OMoa}5%LFsjN@kcw@ogpZGPEY0pm4IbO^0>s*KPvL zkyRdV?*`Z(5%=;uHt7LnOg599y{o&88o?NOU~)9Tkv7qK0mGDro;)a)UiXx5w7GcE`#m_ZExe6ql2U{VNRi!eCM_}=6|LIWS+MIOAE9lp%6#<6bWaJc2x8$&B&-7%LTw!zxO)fPz0P{0p5czO86 zuih7KyM0sm@JB8TS6y`_W1sQJ{1(GKlDR5^VjIH+CY97&=6t}Iwhp|_C!XWkGQV+@@18ugVs(gU*KUVPHY;k)1eb~y8u zXOoz5F5ml>FivCFop;|6F8kJ3cI1@xeJ1IFEqip59qOy5Uk?b3ue|cg@R5&vBvQ6a zT4XkAErU|6j2kzhm&rlWrEjKRPt@sP{m+2UOd9i3B4Stbse;}rvKTz{hNp&ovw5kH5gGY~UgCH32+raNam3ZSu?i))$%%( zhfX}z9h7{n=Jv!K3sVtem`$W&SEwJ5R^} z$>apZhURCW1~<1CLI)Y7XV2}%^NU^t%fg4)#CEbFGzb}X<{TmOLDmO5;MvVZ0O zCUilA5W_DF5^Gz~ld!e|^;nOwXBoreYRR-J06fQ-DRR9c4Qg)8k<^as zm5E^nTsCrjQsglN3xN{J(7rdq9>H@>Hv9p6w2jT=1(2pq17?shd(omf*!D)lV;{YL z`0AHF9d5kohVZs`oJ($!yO|3dl+JOq!1jA1DaabbJXjND0`sMjPN>nbIgHSWyVvJ2 zu4QAUZ7_VS<6N?%V2V_veS19jDlmSKS+n`v^~Fy2z)MeldHD0qH-d;|y6kih)PeC%ciHqf;IP9E3zuDXS-csOepu>9T9+Bf>r(H5dJoilU|;NkvPvA% zV748^PpXgcMdBKDNQl)@sT4#OQ~={Kk|EQ=@ln_GNW^rtDxr8#B83`7v5JCR!ppRD z(fn}Hh3^mh&!0u6Xp&l?q>hek0x(iJd2BEg$dWsQOtU?J(7sL7KE=!H4|gG=V-nmD zlc>^Jc%U3^?pqP&&g-NcQXBETDAQ7)xt9uFs_IZf7+RO1-YM~aDAXjY#&b;bQQj5r z&Aav$6@JT!hs&2Q5C8t}{~oJI)WFrY?^=Lz z^UXI$BlfoU-2PTGT?@ixiJ92TBtDbL*=L^}Gx;0mJe}pfzrtY_v+?LL6l2!tj?OC2 zHP80!8DSQ$##TlE<6A0oBjc`0IEBDCLkzEuI&BQHCmhoZ>1bGTI*@tcxN|-+1N<;X zpz_92XrZugMqo|Hn%#cx91?nUa6qdWn+eY_4!8)Vn=@6|ULxasrPKyfgP45)ZxNb> zCZxhxOStdOwS-pGvg4B)fPL^VGv3=V(shzcrw4FcEm?h?nE#fJ@Px-M3*Y;n%c)-S z*6@=bUlGT66|l`^xL$2;+Uoc1P3iruwVe0f0b)D{bK3}u5j#`XOdRXnN8D%g{KtCd z6#j4C?wh*oN)H^k{Gjl;PkiAahgwtl`ak_4Jng8b@mw~APp0ZBKV;|nif#5l9T;!3 zd-jcuU0q!}m1t<+=#~0i)_b7d1N9!*y**INLXGDH<*`J$q{Rznh5!2er@~>6Jt!vE z(IZ^J1Jr=RDJUvSrv^N$h%Z{aD6CmK2%w=H3W~UN#w7mF#90+;3!BX6zLC*z|H{#r zH9AAGAZx3_z>F_FSQTp#qrGw#{z>WFtmLxGX8c-vqcW^li_IAnZ%QWzs>g($*mrUPJ;)f#SdZO`_+EhHAP{xLI>~W$ zWM1?S42C%vnPx5^VO}?*lChm3M!11Fph~QA9*$ z0O^QJ$qPXH80y6K^M|pefeL?Z$WrUP3xM1xbsIpQhob%Q3qC|ry|csbuDy{xxIb3G#(w_u>4>dmfsqACg z#-$HyYqqk~0korinC40!Ytl(9p%OH-Lu5F37K6Fc(g6tA$BF;eBv~MCK55}X^}87^ zJm$RuOsyYf*Nume*F~h-X!_Y&3J1WtPuY9SVK396Z2{{+Sh_R=etm zI$!}?HN~8IN+thWzK+-Ce59T1TkqE^Ny^4~%VQ9}{TRpot3ne!1Ftm991NU2#jmCe zj!b#p;Hk>6rfl~{u%KIP5BG-ohU>+LCKd$$Eo#h%=FVIT?C|E#L!bR0dJfpvnw(5D zo1J$FJ6$@xwTz{;Iwbjzl)Yav<8Bb?-g&9tlZCd5y1R-noMSjJApp>0%#bdzCx5ZC zfU~RN-8A;FmiZhg8EFF z&MA0|WgKtfI*cj9Pd1CTm(hzBjZUZXXB(u)y`pzUpUd*x&qOU7uf5p*?7FwpHjUbq zR^^ms4&l*20Ox;9>;H_1-!X(JgA{1s_E$!DBXCOG)3^crfI) zA?#~{>Eaf@!5DD`rtc!)dKTy^27UQ4NW2V<+)|O^=9$lGS$XhC)zmOTE~-@2tPT_Z zH+sGi5^P3(Hi^9*0e2320yMv0xbM#4*wvfyPSOQn28f$IF3PlG6Ok|EzY%d=eg6$w9Te;r5Bz6J~ zTiM!xG&-H=dNj!I%4j1A85d*5rg2}sa8W`!~v>A`f_x9`xLptVm-(#w`( z*Ps&L-?)Z+zup@lH>gp;&Gva%N4Q@W^O2uXI6FK;GI&2K+N{&a#`L>?{?Qz=-QoVu z)5c%O5Eo78`F06%O$lCQJor?Cu7ExFz0N(QdU77?EWXdjM~_%MjA~?wpFVawUhXsa z4Ep^$*t^AD#@?#>R(+T%|BhTG_gxCN0g(-a*xlw(zmA%za-zOWe0K1f*0O9d30Ox> zR&Vc0j_sBu8%>EUVPW{Ly*p1HZWo{?OVJ!hTo2Ev(wzh!q--h+NcUXt0tZ-?U6C!) zS+WzVoU81QhW>W!R2aUh3$x!2xV*|FrUiff7GHO11+gpvfyJsM!wP17Z}TVWSEKZ~ zT5#TiZLMcHl-im>;%{lC3d)7sRENZIJxIf#-y&*Vw=A+%>J49xaXA%&1^M~;H@hG| z`wV8A)-D?2{1O9#U6v~PT4$--=W4NB0&B+saHpUv1B=w$=aqd@CXVS__IcgfSy9H$ zmn+;>!*XxAIAmD--+XJu`mP3(j=P?_-@0kF$*)~xRpDuP`)MMPG?;I~u?UFlqT`Ht z!J34k9k3{Vgr^XhEwU$t=!+&(;E}qpUv_$dRYX)~BJahSOtC%yl**9gYDZ;L8jv5e z>U(UWV{mP749C(N3Nv#|B*;v%nuFXr@C|UR={ZeqrW|L%WqZ8Ha`6!LQZh`#(n663 z=Un7PI?;$Sbz1(6;+$@>=5{b;c}_+3WWt7~)3Pm*kbiO^EEypyfYpBlRhN^|mSFaL z8;U64sQfUBC4kA$eiHT5<9&B?E6(pL-yL(@YJTJ24f+QY@B^t5fysF-LZF-Dt_6|5Pe zr|0sdDMjy(8McM)5AJ+s(V<^l{a}YIyiB%cOxl67II`F!OfoUMwNCm}!UGjs>|b3p zv2hNHP1QKJldARd_%LtwDG&jFC~_Y?Jr}Jkw#`gC$4u(?_#0wk^>mtjrK4@Oy#YyX z8P1VS=RN>KvNQrPq#*FPQ?ep>rK9JtsRl5VeZA;b>>55u5ca4TGXIWp z5UB!x5#5FPT~i#d$n`F||3dNIQH&}Ud2`GvB$6ASCO{nB0jGy*!Irt7s~<`G4bFkN zPJWwse>_*wDmgHoj3nt+jMj3-yuD(>2+g%jPG?=yPFCO0y}ae$Dnbhw5)}c;+C3L3L-I#tgGA+w2Mt zJMl@4#8P6)VFRPp<8X%$UbvT0p2VM>^`X0*mB<0)`8vZ&^1YM4)9+9#t3%);2e15+*c*lGL0c{*^ zbsYfHKmO3_Uq)tbLh@WLufN?b>{Sr{PP31%jwdP48imRA&fN`i3-C>6-nsCs&AE~` zb)DzhOf(jKR?wPy)NO0GV=9!hHZNhgt84-^pKDqi#g(>?=g2|nmO00L&)b>aD>Do?j@QF zbp)*Mcwkm)1Nlq88SBL)h0H(XARBS;dUgp2G@R z#@2PN1q-Zk@2M1J50NP}w9WQ0=IQ$wAq*WB7Jo8Y1CBH!&&Lpi2 zIl>pv{V+ONZOhj%DiA}pZFj0y|6mk*tr>T@%RV};Gu>nbIqcm_tbCo@(X~K%1(H>5 zi@E?ISU7Z@hbS(6ul9ow8Ny%FeZD|ZJ|9@6F4?|ruCfobfn?5+KmaZKLR%e?BM?m* zo^#bzwe1B&8QYO+yz8+LxX$?f{`L&J+~|I~bP?m=LuR#Ar}=d9wIA;AWvbLyA;uRg zh36NWE<1|i`5nxiLtv%u^#wxu*q3uEH(+-iQ%>MLF0iZKnJZsXix_v<#Uxjp1J3eM z_?*VoCcrhenw=@@^R*Pv?cQK6RNe-jo0wgwcO|QVLxE5?v1oz{UWjE|Ajqtra+}b? zDOxdia+*MtG1D^bC-SJ2XKiV=oQs*0nr9Mi^hFM{K0bVv_i4tazrQ;>^AtOW@D!td zBSYLMAyaanR|`XHCoE{Mxxcr1jRyI2O9_8RhnjISbUbY3O>M>5>bY0Ic3gaC6f3&= zu=6Xk4gV~2#VD(ho$aUwJ9$$r_V02eNDWYR znB;h|F-M4k>4AMc4hwzylMj78$kB>0@!1efNSor?RSj_4z(NGaw5o6=)D!$!3qxBvpFI;#EW<)jSMKZUWO*;c zC2o*G)hrAc3h@Is1cbfBd3r?T4X|%VOqU!2F(5c1-#VPiD%z8TWB3Vz51Fil*`lM%V0ySS!(xtg6~X) z%7rRHb$KHF*ub4L$>A;!djWUZ)^x*fWX|R4r-^1CQHT+#J_ETLWi(4|f2V!dV)BsOZ{dHrL(O($9E+(c*J`*Fm85Pi2P_x!kq z@kVGyg|b;$O)~kL>ueLo-!c@*>o1r7JPCc|@?dI)bjg4of}t$op7W@AS95{xv64Y;7@T$vtC-F-j{TtHrN zR9eot6%D|seVw+d%{ht}BVNySQhfUuX~J`Ww_%j4+~^OK9>hZFuxjqp%%$o_Fcg5# z+uNJ$^wosuS(6JCZTMX161-BLUy2IK%22Zzq5@}3%bxL@>P0Q|;?!;S-Dgo(QzC2&g1t7SQe1*Cy<5V?Uzm?A>M0R`{6Z6(H32ttx`Ck4rN! zC^yND;Tv*Hqd4pye`x7!^@|S`J8iOA;xLuxlTI%tnS0ES0K8d$4 zKIp|&-!X-SjN4P?#i+2lDE*c+aRw@IrB!&2){q5TF3dQc=Afp$cXzzS?`pPZRlBnk zWjK3g43nd)?4fUGQQj29k0knf5T~OXVdn|h%mI}kxXLfK-=G8L+iZJ8$+4#KfFknm z(UY)%u^WNqGZu?CxlV|8&+$y@NIn8mU zoFPrq1w#A+;U0}@gbTM%;D8@;6X&YGPo`cW`e{dXjiRq?e(6TOb9a;SrNM!j)Ql6>*Rk| zT5tu+HY?$NR~hpxvVZt1CCCcz(j~-Eqx{9ir`+(ziXU>X0^)DBN-v=CVEDZO;;dAj z(~8>rhgp-o-|ky(@eV@DEgdyHh3cIL@-VI4c9auood;5bEb00>UANh9%;+WX8efy} z?3=(xSTV}S@*%t*Hp3(x?P^lR)Rc+_ip2j?p>&=$EN{iN&oh0QC%P8eQFy zG1GRx^_$)=>0Ec=j|VfZQt~!{45O(tHY-h!B-G|FSGIwA38cN8azr|pO@M$GPZH%1 zTmHhRiIFl5%4`tDn&?N;Kc2G{Vy*&>B3;xas)9Pt%PpqiTC+bDnH*)Ep%40rWo(TM z{nyjzHYMX((^y{^v<9(9zmpEN+@P^5n-A^K5oD`SgYW)rUPhFsS~p$lD_{Q8qU)Z% zV7L{b4EQmqwA^WBVw)SlKH?VTNp`cPPE$cOw+k0Aw&jgyGT^kGu4Xj;Ij)zhv7`Q2 zd4?WyI#9Bitbr02(bCu6mJN`>j&^`L;fe2(0N<_+brHa9LQ-7LB}fH2D_h^7@G5fq zf1t96I(I?Qu4I-N*;F5-_mk)Kmpx8&&eItSWZZtl&Qm)FM0lz@C_ro{NQkfhQ5B(b zy*!bFNdjzuXMgID`~m|zx!x&*q)?dJIR45{7R)FOH{FyCn?ng^p=E8syM0g|Y(*xa zMh+WrbMOWHPpnqoI;AX6hM=jdX$ai~s|b4De4$qXkk&cNceUFSZw%Qb))xCGp)>2e z7E>edDCwy2Va@n&ns`M+ImZz?B@*>?a|9rXcQ0U%U{(I7g}a4j8k{T|F)hjJ6J_E zJoX{(c&CrhqgVsQ zyc~uB<9-d3TlGA!a=GzzX0o>13J2toRC>gZGo;BQ=I6DPXM#@4*D=ELKwRuf>hU4( zEGBkb@{REYq0+Se%}W?e)c6R75)Et&oonpvQ)OlfjZa<5Po{841_nobi2J^jeXqAu@%02`MZ z#~7gzR&RS%!hjeJFVAfhur_|a0FQKD1ld%Eh=IE4FG=2)lr#^>Dq^fN&a>Mq;Tj0? z9hhAXo$#qHqGZ+UxJy4n%2%iTRaByGjIcJkxK0jfzE(QtG`=LbB)O!x7~WL6Opou5 zD?_EvQNIvmln!T0z4Sh%QY= z7Yhw8LNiM{3 zl+EfR=^U&b*_7sBd1)t&OmzA@f^?c}f}=%+Iv(DiVXEDi&iPS>!8BS~=QzeU9tn|= z|CUOoe{Q+JU3WV7{fX0~hGpWqz2sMNL$Dcm+cW7o3$fk%VAfr8tzI;UKj2-u2p(Xo zqnk_qSWDzXUH@3q4~%%-POk#x?sR#-ZMSct9APxp=Z%~S3GS5`_|Efte7_cSK4Z|U z@H-OK4Kz_$1tK#p^=beIWo5&8~Y#TdU+CQMb0URjNZM44*6`>iP z_*Hq0*^tsRaDN5whp^aQL1PRE&}FR6yLQQa)?Lwa^xBG`lz^7kF0^|F9dA>!XaMw3 zH4q|!wOetyk;#O{fh99tP+W*2(^_iKeddDS)X;JtM=ko7vgQr^eYv*6$w&=veb&hvKktTE8gIY^z{h2QcKHV-oegt~hl_JjmZHNqD)ntj!(Y!!KqhW=SyL0jkM4p~S2heZN!d9S+4 zDZ0(!kE|M@@Fd3^Du8?=Jxvp=Wb->@j4&80^YFV02-v(8gX8Wy$6LhCt?o<#v6LwtBxpu zPjp7Rs&?-AoVXv7vpW$rz(4gqll+!#Qv}3_@xm3J-dfqw}7l_FHu^)KzW-^}nR<#qU^L6qK>$^gN zImUeL*00UFB(8=&wE;5{IL4>7J*|un!|&^B%>(^zsa7xb{2hI<*}ctd($hRvQkF6^ zX0;G4b`WY6Wp>(PKQU-zEFr(upQfg?Y&J+Kp+tpHy+D_tkaiKVp*2>D(H6cIAg9qw z$u>%9E&Y;rabRfJUS7zl9w0oweh!`B&sb2vUmrM6i=G&5_TA^(&pP*+xNpPq-c^dq zWVtFUw-e1&pL1SfjMyR_SkgTuKs|+D)2O^l;8B{b&iV-^{j(x3r*Y6(vqaEf66{T* z<0a+6Wn3 z)|AsIEIRI@7rPB8_vK!m_jt!R#^Uz)v$FJk_3^Mcs7?FRJa<{L54kNO6Ry#N~wbJS0_wb)mmBO&<`^j*1U$?ZBB$d94(G&qc$)GZz zqwjST#*O*5OCcz^Azw4;C+8A&1Z&4WsiDGgX3x}_)lD!wZZMQqed4xpIb!0ZE34g& zL|L9&vo6=*C}qv8c2O~69sW+EBd1g6+MKV>srkBvN!lx)s$Dl4ifNYl$*q%TJcS*Z z^w6KRxN@fY5PJuL4wM%cS2^i)YvFw(9n}^UIilO5pOBBUa_F^xd)LYM(m2}eRG&Bf zqlTi!y#%AcVTR2$xn5vjyHXgO=YcB9^{!wE#=SW{Nm>&wo~Km%L3e-l*ClYX78?d9 z1bx|1UU*rpz&F*oTMc&}mChVH#8l`KJ#^#na7Z&J3U`hjGVgcqPL&C^ zk)}rPCElVv^SXxAC_0v=l;h+2tTQX1FjJhpt>ItEU?srQluBarKyxV4ix8R%X~W8~ z9_e4r?*Wg_?E_k=bLVFAEP?jgCVKpk*lao#Fk?sciZ6Mvw~n7>ghMebSoog`?YNe| z90s3O3nCKO3a49M}9^c5>)BM;Pyna{0eudZFtza3@JCY8^ z#Y~8fYp?+)M+8x5o?i=TwAFVxNEWw%1@c=aASgQz*<{_k#~}8UxW6Q%$Fd7PHC0jLGK^Y_-Jqi!RG*o^AFZvJF@Dh-97kg4=W<9Uu{(}s4mvsHak z51Y0p(s5C!XL%3+w1<+)ZnUuCv8{DJe_4i2vd<#eww1~@8>+;V4|;xoKo$$ji4Bq&C8mRn9s}Gf zF`d5kb9SHtgR$FskKj1bfdHi%lg%dy+AJ=I4n-Fgf1JFnHv$yl9RZ|nFjxO$=EN#6&GuspCSazodKJX`0Bg0*tB-84k(QdxK9s zS2(jV=hZ=5yDT+pN}~xRzGCXUtvrQyzbe$vW$Y-XavOpj+A2AI4)G{|w@S9`6uX#S zdYifA#B8!DLO$Rdcg>nH%-upQ854dT>kgO=J(&CqEzA|C4dh_B*kD2KV%DefKxSw7 z6LfGzdP#Z}2zPvoKqzV*T!KCae1Ga;_H4$6(0yOFe=Lrw;MlvuTUT|Ylug{6medA% zjc6}lLcY9M;Dcj``xwA?puv9;FfV~m$OCan68m0;z6{#j zM1X*pvcRZyD1B+P<|c*9WaW!$8D|`F<}X)<{USnTQsGHEIO4m;6w;(g&z%eI>GTCN z0{1H}-Fzr+nd636O$cNwSj*E7Bc+h_f?qI#G>LDkQ7)ers!h80xyJGA0L}ZnOC3II zMI-^c#5)>)qIi4@)f1M={qTj$Ou4and|aPs%7*O{fMKwgt!# zX}Yyyk%xSrMi(GV$<^fCc3fEf~yCCx^A}r~SxtMc`R983Jh$ zUfTtlIY;#=Hn^w5ZW<#u%E4M!WM}g);H4t%SMrs>F=|PtY=J`UGGtRc|1AVJsW6!7 zqm$vz#^+?7dhTQO5Gzn3d?$_SV3yCPT9v5qEO}p*=E5i2wzJcQ%WNG}h9yduv*@!! zTcy@eQvt2ULS`fddaJTpd+6pKsJilURfOj7wz;qu=QmMdr^Sguxv#C@%by;Ule;C2Z5 z;?HZ-vjoJJ@lQW+q((HjFoTASD1^Jp+Xh`ab80s)^y3*X^ZV}Z=ygu*<~yfTfy(wy z%~GzPhG7kqD@JwFS$UP!168!YEw5pB^yQvdQVA^Rg7b@whxWH0I*bKkc(0NDu z{F;V&9@mnj4Sc>CR}MJ__idS|{~ll{57-MQj_YI&Y}7VY&?^T_xV!4S*38Z^in0HYt&JG3^+T zMo9>cMRmv`fi6_>CgJsIhvJ5GEN$1*=g9X5H**5-_h-GHZw8`*!Cs+VYwg~mc}XCp zGLGocfgYkmx;HzRUyZ85(I98SXV_buD_H7nNG)2A15~tm)m5mLJ4b~ z{_lh551@g*-By}4g&Oi;@)V=UvKJYB@_t+8pIu|Eq@4(zh$4H_9^1*RM~R)5#vKag zYe0|)g*ViGv=&LVMHwYBdL3v?TESW(MMYhDB<%i>9joHdhKM%QPP5UCm88s}M{Z9Sc0R zknRo7bIlWlU$)TBm#D5|CZEV;@OFCLgwuc1OP9~JNi*v+cIyh(`b8=kb<^PCet^BR z4+IR82M3E}W7w0P1u3dS+NV3+(dMat4;Pg%-xFQ*3GdP=ngmzlctLqi$w4SbRx?!@ zG6etQGQ~DvAFLvXKXW`e3=(OYL?MsPi)`}VI8s%5Wus9uF46(tq(U;!KVjME<6A@I zrYRd=4N>5A`Qui_0CpKlGP5U%} zt70^h7AW~R%-UY_&=sah!&2rh!wgGfDqUkc%QNe=GzF-fRw<@=SY?Spj)aeRl?#b#$vNG=l-bnVg0DI5tUvz?touESM^# z$q1mq?<^x6>Pm#uq7x}GPbVDdQhK^Dr~*f_FQSp`2NG&_K>0zeQ?nyfM;1P}bROm$ z9y(ve*e-F@0Af@(T+&D_ry7(z_|7fE&_d{V4pNrwcg_rvjZ%>Rl1424>lM}4k;2a+ z$ayRN@2{btd_&Bc15M&MD9ot68K*Z_#ZHjlYH=UrvAQqkcj_5`#=rdsb%|sM2Fg}QL=0j`dUmjP|OrtPNpcGcj5Gmhf#8GxR-6JDkv2{&y88dAe;Lw9DKL15MUV?Vi>WPo*l) z<;SNVE4}e=fM!^6MDnwJ)4y~}jzOx-o$&)26NzX3z>drt{o@2~mdT&xSk%Z%TiWGc zGJq-<^fR*v8N&B&MJY0+i|P)xsc+@rEU(8ZO6|hbO_+fGnk5G}8HD<&ue&pqGxMBf z{0Y`sN$~YS2c_LgS+sKH?3mPgB3-5{%BCyn??sa;eceizx|aZBPOTLnUmuwv z-Z1sFcBEU2LfCtdP0;bx6Zmuy6gm z<+w5)78?SInoNUn&*`Do$eKUf45XY&)>;fH;yP`bmo{Cj7$(FDG9xAi0qC!ZY`n3M zh)zU?1k%jc;R4yZf|Y4mU(>M88JqR5c)Qz_J{ezBm?D%G;*bVINXsQYU!2!Q35Bba ziweciL(S$}l$@*RdCiH%1v{%$ej+Y&Y@WfVN6y}B<0(E`huKXxJK`=g35Uo=Z&nJ^ zu4U>bL|c|fp_v}mdL-O?0)Ps>q2G-}qfN4Td7gSCw9%!TYHjL&^IKkrHrPGOx!d3^ zA(!Haw`%-UT61{jr_*6vR6ktja;pAF{pn6qV+1v#PjAlSEQDuY#R^kgW7F=`|Io<#Kt}R}oMoT4zjHR_9oACOhht(>*F{^$3jJqP`pC?$zp(&rms^ zFdl9(w|WR2&c))5w64FN$?^!(rAWPN3{%kdC3z=O+p!=pbH&V9)hpFH=)PN`b5CAg>%Ly;RRm`i1S9^|y-&bj{w5+_nvW zZgJ**D*yIr4Mi9rG@;tA%QoJYkp}M>7v5nEd4Ns4qG!$d zRAW6|h@Z=|YH#%@lFL&b$d(~mY+WQYLIg+%zx%z^W{rGu8ztsZhi)Jg91;y%#$Q@S z_$lytShd&0&{2t8-vGw#A33o4LB$$Y_E@SL_)b)39>$*Snn1&^Bym?Z{iV4LXioHCL*K9;BJTS^VSKIeC&YkBFjf(H+0tt;MR3d{qBj6`L)C-F}T+QECXD7kCch=zf|(;f@xu$Bd!5}wyuO{_s7fmFfN#fYZK?}v#}vJ*=nfd zU#Aa?(=xAfe&AxWF? zq8z2jS=fpQ0_zK?O>9s~;^90oUTok3{WOJ4BJwh5N~G%VAler4FMC@LP?FTLA}AW+ zH)4=$!gb<7K9sA>7a#8-mzS#%#<=*?4}f!n|NC2zanYpiXcoyK6+8$+H8!{>-Dzkp zLMnajdUG(D;PxP5Z-}sGF=U2vE5#=LIp~o^h8da#B!tTOedP}nFz9CBCB1JS#b1;z zsZ8baG7#Q;ALGS?WV_Inl!#KfGzq#s6K)9#T$Chl!(r_49i$8nXO-2zOthbQGr6rv zycQjsbQK96NhTS)Fs&#^07FMS%G`&k>4?gdI*LAws5b!7Zy=bot?v383!b?^a0SXvA8G5uFyriV zL0QXErYU>iv9S8_z3VT;()1gKlH?_st`0eW^-73g6T0bL{9!vq2?jK011XP@SaJtD z`r-HBn9Jjr;NP4+6GcBrMvBy1&az3GPPOi47D3KM)?vVdX! z?B>ZB$qvc3KH2ac;$cYBXH9%(h}Y)>ME)yE14+ANM33SBU{JM68t~Y`xY5+CWem#+ ztT-a_U@W9HEP#m($iT4!d={d%OYdK-;Q)CY5YuUNsl}~4JL%B6#!y1u)K1MUb=70t z*N7=--L)fpLALz?Fu(NTsyX!@u=Tr@R;l>IWc}DqYUMIhJT86rx(!$39$y;a8ZCWRe_Jv4=o9v z;yk`%&hJ!zH-viKRA`{1sNqYD^v3oLHuY%=J&nOq)*m3Y6IC@uF;oXaMfl|aOYCQ)aRC0wJlY) z)Nqex$ol8}$CVExoU$~1$UL53eDWAH%_YXHRnqtFKiR0F)VTa`*kiV+CoNAR*`4-! zh6dbO;|G>XfPG7V8#xEew@M7oS-LS@#Sk%P=UgHbCL~-I;#_(j8;r9^9mz7bV+_&+ z>J!JPYAp!%`t~6}Gf>m)Seplc*We-&U^(K`V$XR#M?GWPAf#r|nt^=sMvg{6H{TA> z^d0>SwMWa`@%D7wBV@*MC(OKs^!x21Oc04E$WhNe<8$wW^Tb#qtvr-@6!wHTOIF(- z!~|?byDGMbU;Z9Y_<>g9VB)D0eF!1(sFuTbr^~_X^yJv87Ch4FRwhR=Dp6LflhwuL zLNdxu#pS=>C#9Hge}-US@idzx7|#(n`1d{2ydlH;gF=Wk_?G>7Az}l_n6}}_r>|qa zJ%Ti;|Kyj(E(ogo3aq1QhQmcmgf_NW1bgLdHal^0G3VgTcgOufP3M3j^nY>k2HhS8 zSwDUlFF^pNz>c$)V+bRLyMcB?v)oP4D)mwFIm!bljv1hjAF%_Eg7mw&nOl>HId=i` z-&{OG`8ZH%D%|IQR&vfi9sZJrsiRfheA4`{UW&Avn2uL4m&at?S4g>S83> zRpvppNYxtWAH{cNNdyqbq-9pZ8<;YFRK^WO++z_}zqE+He{z%X#b&K$P8TN!{}*d@ zJ^FTr(Bd&@CZK``>HmZ~S#6y^>v%pxc?lj|NLO@N_)SvMHzD}$!&IBo*K<=w04~`D znm502z#+i!We1+*l7IAi1M-Ky4#_(QYki_!WgX zeeOdy>Vkyj>ts5-QoY~n(=~+hSZ}o_0q7w=mhal&t`A^eJi~af$=IhieH$y%4b!H& z%5wDWsg@i+j@YgyxOg;UP(H;b>se&SU(l0)Zj|y&QHb@XtNd=u<3%jnl3G9%Nm1hd z5Af@D#oh>hAaqPX(&kkAOHO5}YgI*V#=olm)hI^-QG_2B=)&uI5YPX5Gu>b@kuxGO zhIu#=I?kCz0mVR}H*IO^8GME0klx4FQwyabt#KUbBy9j9$Ng&uQ0*Vh(JyIuy*GB3 zeeOfqn~m)+Z*T-hhYuw|)^ot`!6c?52W+o!$-s4KPe97cA{@5cJ}%O{ik{UeZxSpw6(4K*tPF2dOqXHRAYh4 zW%9oMn{ll|i>c6i_#|713zluAv`Fz$oA+@NX`-w=v7PQv) z)xI{8{U$~^us!DBOwvb51FR&u8uKym{cr1jydQqxV+0uYiQvu8&(I=mzR!1v3i-9Q z(|#aX&&4D~0jsLcx0ctB2r=bSM$JhJE(nU~uHyG?d(I?=&12pK)$5WObm5VCw~yZX zpYouix_lyJON|NErs_W<dYUe=N(Q9?jPAfYMgeCm_5!6jqZfvcmtc*H%Jva*Q z=VZ($+~wm2>Z5d4xF+h`tY!>9;{ScIkN1<~hX8>sBKcn~db&rSK=;nIoNJzBY9eL7 z()0mC0w6i*#jOnc(1HR!=DXlz-K_dc8F(nbkD z-RF6xiztZp-

14r)bw0~hVe)O|B?h^RtJO#9qbdjSt_IaXhJpTIcb9}6K|M}^f z2=t)W@3>Dm|M#o^|8sH>pg$*pO>q8q{`_AnfO_g3?nTg075h&2a&Z$3^d&8>AXX`2 H81(-D#ioNH literal 0 HcmV?d00001 diff --git a/pyproject.toml b/pyproject.toml deleted file mode 100644 index 1c14a08..0000000 --- a/pyproject.toml +++ /dev/null @@ -1,56 +0,0 @@ -[build-system] -requires = ["setuptools", "wheel"] -build-backend = "setuptools.build_meta" - -[tool.black] -line-length = 88 - -[tool.mypy] -ignore_missing_imports = true -pretty = true - -[tool.ruff] -lint.select = [ - "A", # flake8-builtins - "B", # flake8-bugbear - "COM", # flake8-commas - "C4", # flake8-comprehensions - "RET", # flake8-return - "SIM", # flake8-simplify - "ICN", # flake8-import-conventions - "Q", # flake8-quotes - "RSE", # flake8-raise - "D", # pydocstyle - "E", # pycodestyle - "F", # pyflakes - "I", # isort - "W", # pycodestyle - "N", # pep8-naming - "ERA", # eradicate - "PL", # pylint -] -lint.ignore = [ - "E501", # line length violation - "C901", # `function_name` is too complex - "PLR0913", # Too many arguments - "PLR2004", # Magic value used in comparison -] -line-length = 88 - -# Ignore import violations in all `__init__.py` files. -[tool.ruff.lint.per-file-ignores] -"__init__.py" = ["E402", "F401", "F403", "F811"] - -[tool.ruff.lint.isort] -lines-after-imports = 2 - -[tool.ruff.lint.pycodestyle] -max-doc-length = 88 - -[tool.ruff.lint.pydocstyle] -convention = "numpy" - -[tool.pytest.ini_options] -pythonpath = [ - "." -] \ No newline at end of file diff --git a/requirements.txt b/requirements.txt deleted file mode 100644 index e484c81..0000000 --- a/requirements.txt +++ /dev/null @@ -1,10 +0,0 @@ -black -flake8 -isort -mypy -pre-commit -pytest -pytest-cov -toml -types-requests -types-setuptools From 13d090283777f55d6d6e072f755ce8b30e3ee5bf Mon Sep 17 00:00:00 2001 From: Sana Ayromlou <46001482+sanaAyrml@users.noreply.github.com> Date: Fri, 7 Mar 2025 10:00:48 -0500 Subject: [PATCH 7/7] Revert "Update solo library" --- solo-learn/solo/args/linear.py | 8 +- .../solo/data/classification_dataloader.py | 109 +----------------- solo-learn/solo/data/dali_dataloader.py | 73 +++--------- solo-learn/solo/data/inatural_dataset.py | 77 ------------- 4 files changed, 15 insertions(+), 252 deletions(-) delete mode 100644 solo-learn/solo/data/inatural_dataset.py diff --git a/solo-learn/solo/args/linear.py b/solo-learn/solo/args/linear.py index b498d7a..35330de 100644 --- a/solo-learn/solo/args/linear.py +++ b/solo-learn/solo/args/linear.py @@ -21,9 +21,6 @@ "stl10": 10, "imagenet": 1000, "imagenet100": 100, - "food101": 101, - "places365": 434, - "inaturalist": 8142, } @@ -33,9 +30,6 @@ "stl10", "imagenet", "imagenet100", - "food101", - "places365", - "inaturalist", "custom", ] @@ -169,7 +163,7 @@ def parse_cfg(cfg: omegaconf.DictConfig): ) if cfg.data.format == "dali": - assert cfg.data.dataset in ["imagenet100", "imagenet", "inaturalist", "places365", "custom"] + assert cfg.data.dataset in ["imagenet100", "imagenet", "custom"] # adjust lr according to batch size cfg.num_nodes = omegaconf_select(cfg, "num_nodes", 1) diff --git a/solo-learn/solo/data/classification_dataloader.py b/solo-learn/solo/data/classification_dataloader.py index b8fd07d..fc44333 100644 --- a/solo-learn/solo/data/classification_dataloader.py +++ b/solo-learn/solo/data/classification_dataloader.py @@ -27,8 +27,7 @@ from torch import nn from torch.utils.data import DataLoader, Dataset from torchvision import transforms -from torchvision.datasets import STL10, ImageFolder, Food101, Places365 -from solo.data.inatural_dataset import INAT +from torchvision.datasets import STL10, ImageFolder try: from solo.data.h5_dataset import H5Dataset @@ -137,81 +136,12 @@ def prepare_transforms(dataset: str) -> Tuple[nn.Module, nn.Module]: ), } - food_pipeline = { - "T_train": transforms.Compose( - [ - transforms.RandomResizedCrop(size=224, scale=(0.08, 1.0)), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - transforms.Normalize( - mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD - ), - ] - ), - "T_val": transforms.Compose( - [ - transforms.Resize((224, 224)), - transforms.ToTensor(), - transforms.Normalize( - mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD - ), - ] - ), - } - - place_pipeline = { - "T_train": transforms.Compose( - [ - transforms.RandomResizedCrop(size=224, scale=(0.08, 1.0)), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - transforms.Normalize( - mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD - ), - ] - ), - "T_val": transforms.Compose( - [ - transforms.Resize((224, 224)), - transforms.ToTensor(), - transforms.Normalize( - mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD - ), - ] - ), - } - - inat_pipeline = { - "T_train": transforms.Compose( - [ - transforms.RandomResizedCrop(size=224, scale=(0.08, 1.0)), - transforms.RandomHorizontalFlip(), - transforms.ToTensor(), - transforms.Normalize( - mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD - ), - ] - ), - "T_val": transforms.Compose( - [ - transforms.Resize((224, 224)), - transforms.ToTensor(), - transforms.Normalize( - mean=IMAGENET_DEFAULT_MEAN, std=IMAGENET_DEFAULT_STD - ), - ] - ), - } - custom_pipeline = build_custom_pipeline() pipelines = { "cifar10": cifar_pipeline, "cifar100": cifar_pipeline, "stl10": stl_pipeline, - "food101": food_pipeline, - "places365": place_pipeline, - "inaturalist": inat_pipeline, "imagenet100": imagenet_pipeline, "imagenet": imagenet_pipeline, "custom": custom_pipeline, @@ -271,9 +201,6 @@ def prepare_datasets( "cifar10", "cifar100", "stl10", - "food101", - "places365", - "inaturalist", "imagenet", "imagenet100", "custom", @@ -294,41 +221,7 @@ def prepare_datasets( download=download, transform=T_val, ) - elif dataset == "food101": - print("=> using food101 dataset.", flush=True) - train_dataset = Food101( - root=train_data_path, - split="train", - transform=T_train, - ) - val_dataset = Food101( - root=val_data_path, - split="test", - transform=T_val, - ) - elif dataset == "places365": - train_dataset = Places365( - root=train_data_path, - split="train-standard", - transform=T_train, - ) - val_dataset = Places365( - root=val_data_path, - split="val", - transform=T_val, - ) - elif dataset == "inaturalist": - train_dataset = INAT( - root=train_data_path, - ann_file=os.path.join(train_data_path, "train2018.json"), - transform=T_train, - ) - val_dataset = INAT( - root=val_data_path, - ann_file=os.path.join(val_data_path, "val2018.json"), - transform=T_val, - ) elif dataset == "stl10": train_dataset = STL10( train_data_path, diff --git a/solo-learn/solo/data/dali_dataloader.py b/solo-learn/solo/data/dali_dataloader.py index 406b56b..8451846 100644 --- a/solo-learn/solo/data/dali_dataloader.py +++ b/solo-learn/solo/data/dali_dataloader.py @@ -20,7 +20,7 @@ import os import random from pathlib import Path -from typing import Callable, List, Optional, Union, Tuple +from typing import Callable, List, Optional, Union import lightning.pytorch as pl import nvidia.dali.fn as fn @@ -32,7 +32,6 @@ from nvidia.dali import pipeline_def from nvidia.dali.plugin.pytorch import DALIGenericIterator, LastBatchPolicy from timm.data.constants import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD -import json from solo.data.temp_dali_fix import TempDALIGenericIterator from solo.utils.misc import omegaconf_select @@ -203,7 +202,6 @@ def __call__(self, images): class NormalPipelineBuilder: def __init__( self, - dataset: str, data_path: str, batch_size: int, device: str, @@ -222,7 +220,6 @@ def __init__( are normalized. Args: - dataset (str): dataset name. data_path (str): directory that contains the data. batch_size (int): batch size. device (str): device on which the operation will be performed. @@ -247,59 +244,17 @@ def __init__( self.device = device self.validation = validation - - if dataset in ["imagenet", "imagenet100"]: - # manually load files and labels - labels = sorted( - Path(entry.name) for entry in os.scandir(data_path) if entry.is_dir() - ) - data = [ - (data_path / label / file, label_idx) - for label_idx, label in enumerate(labels) - for file in sorted(os.listdir(data_path / label)) - ] - files, labels = map(list, zip(*data)) - elif dataset == "places365": - if not validation: - split = "train-standard" - else: - split = "val" - _FILE_LIST_META = { - "train-standard": ("places365_train_standard.txt", "30f37515461640559006b8329efbed1a", "data_large_standard"), - "train-challenge": ("places365_train_challenge.txt", "b2931dc997b8c33c27e7329c073a6b57", "data_large"), - "val": ("places365_val.txt", "e9f2fd57bfd9d07630173f4e8708e4b1", "val_large"), - } - def process(line: str, image_dir: str, sep="/") -> Tuple[Path, int]: - image, idx = line.split() - return Path(os.path.join(data_path, image_dir, image.lstrip(sep).replace(sep, os.sep))), int(idx) - - file, md5, image_dir = _FILE_LIST_META[split] - file = os.path.join(data_path, file) - - with open(file) as fh: - data = [process(line, image_dir) for line in fh] - files, labels = map(list, zip(*data)) - print(files[0], labels[0], flush=True) - elif dataset == "inaturalist": - if not validation: - ann_file = os.path.join(data_path, "train2018.json") - else: - ann_file = os.path.join(data_path, "val2018.json") - # load annotations - print("Loading annotations from: " + os.path.basename(ann_file)) - with open(ann_file) as data_file: - ann_data = json.load(data_file) - - # set up the filenames and annotations - files: List[str] = [os.path.join(data_path, aa["file_name"]) for aa in ann_data["images"]] - - # if we dont have class labels set them to '0' - if "annotations" in ann_data.keys(): - labels = [aa["category_id"] for aa in ann_data["annotations"]] - else: - labels= [0] * len(files) - else: - raise NotImplementedError(f"Dataset {dataset} is not supported.") + + # manually load files and labels + labels = sorted( + Path(entry.name) for entry in os.scandir(data_path) if entry.is_dir() + ) + data = [ + (data_path / label / file, label_idx) + for label_idx, label in enumerate(labels) + for file in sorted(os.listdir(data_path / label)) + ] + files, labels = map(list, zip(*data)) # sample data if needed if data_fraction > 0: @@ -1048,7 +1003,7 @@ def __init__( assert dali_device in ["gpu", "cpu"] # handle custom data by creating the needed pipeline - if dataset in ["imagenet100", "imagenet", "places365", "inaturalist"]: + if dataset in ["imagenet100", "imagenet"]: self.pipeline_class = NormalPipelineBuilder elif dataset == "custom": self.pipeline_class = CustomNormalPipelineBuilder @@ -1085,7 +1040,6 @@ def setup(self, stage: Optional[str] = None): self.device = torch.device("cpu") train_pipeline_builder = self.pipeline_class( - self.dataset, self.train_data_path, validation=False, batch_size=self.batch_size, @@ -1114,7 +1068,6 @@ def setup(self, stage: Optional[str] = None): ) val_pipeline_builder = self.pipeline_class( - self.dataset, self.val_data_path, validation=True, batch_size=self.batch_size, diff --git a/solo-learn/solo/data/inatural_dataset.py b/solo-learn/solo/data/inatural_dataset.py deleted file mode 100644 index 65b689f..0000000 --- a/solo-learn/solo/data/inatural_dataset.py +++ /dev/null @@ -1,77 +0,0 @@ -import torch.utils.data as data -from PIL import Image -import os -import json -from torchvision import transforms -import random -import numpy as np - - -def default_loader(path): - return Image.open(path).convert("RGB") - - -def load_taxonomy(ann_data, tax_levels, classes): - # loads the taxonomy data and converts to ints - taxonomy = {} - - if "categories" in ann_data.keys(): - num_classes = len(ann_data["categories"]) - for tt in tax_levels: - tax_data = [aa[tt] for aa in ann_data["categories"]] - _, tax_id = np.unique(tax_data, return_inverse=True) - taxonomy[tt] = dict(zip(range(num_classes), list(tax_id))) - else: - # set up dummy data - for tt in tax_levels: - taxonomy[tt] = dict(zip([0], [0])) - - # create a dictionary of lists containing taxonomic labels - classes_taxonomic = {} - for cc in np.unique(classes): - tax_ids = [0] * len(tax_levels) - for ii, tt in enumerate(tax_levels): - tax_ids[ii] = taxonomy[tt][cc] - classes_taxonomic[cc] = tax_ids - - return taxonomy, classes_taxonomic - - -class INAT(data.Dataset): - def __init__(self, root, ann_file, transform): - # load annotations - print("Loading annotations from: " + os.path.basename(ann_file)) - with open(ann_file) as data_file: - ann_data = json.load(data_file) - - # set up the filenames and annotations - self.imgs = [aa["file_name"] for aa in ann_data["images"]] - self.ids = [aa["id"] for aa in ann_data["images"]] - - # if we dont have class labels set them to '0' - if "annotations" in ann_data.keys(): - self.classes = [aa["category_id"] for aa in ann_data["annotations"]] - else: - self.classes = [0] * len(self.imgs) - - # print out some stats - print("\t" + str(len(self.imgs)) + " images") - print("\t" + str(len(set(self.classes))) + " classes") - - self.root = root - self.loader = default_loader - - # augmentation params - self.transform = transform - - def __getitem__(self, index): - path = self.root + self.imgs[index] - img = self.loader(path) - species_id = self.classes[index] - - img = self.transform(img) - - return img, species_id - - def __len__(self): - return len(self.imgs)