Add running scripts to dali

scripts/solo_learn/barlow/repo_setting/eval_solo_learn.slrm

@@ -0,0 +1,30 @@
+#!/bin/bash
+
+#SBATCH --job-name="eval_moco"
+#SBATCH --partition=a40
+#SBATCH --qos=a40_arashaf_genssl
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:a40:4
+#SBATCH --ntasks-per-node=4
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=24:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/imagenet/ \
+    --config-name mocov2plus.yaml

scripts/solo_learn/barlow/repo_setting/eval_solo_learn_synth.slrm

@@ -0,0 +1,30 @@
+#!/bin/bash
+
+#SBATCH --job-name="eval_synth_moco"
+#SBATCH --partition=a40
+#SBATCH --qos=a40_arashaf_genssl
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:a40:4
+#SBATCH --ntasks-per-node=4
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=24:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/imagenet/paper_setting \
+    --config-name barwol_synth.yaml

scripts/solo_learn/barlow/repo_setting/train_solo_learn.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="barlow_train"
+#SBATCH --partition=a40
+#SBATCH --qos=a40_arashaf_genssl
+#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=96:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_pretrain.py \
+    --config-path scripts/pretrain/imagenet/ \
+    --config-name barlow.yaml

scripts/solo_learn/barlow/repo_setting/train_synthetic_solo_learn.slrm

@@ -0,0 +1,34 @@
+#!/bin/bash
+
+#SBATCH --job-name="barlow_synth_train"
+#SBATCH --partition=a40
+#SBATCH --qos=a40_arashaf_genssl
+#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=96:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_pretrain.py \
+    --config-path scripts/pretrain/imagenet/ \
+    --config-name barlow_all_synthetic_icgan.yaml
+
+wait 
+
+cd ~/projects/GenerativeSSL
+sbatch scripts/solo_learn/barlow/repo_setting/train_synthetic_solo_learn.slrm

scripts/solo_learn_dt/cifar10/barlow.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="cifar_single"
+#SBATCH --qos=m2
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:4
+#SBATCH --ntasks-per-node=2
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=4:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/cifar10/ \
+    --config-name barlow.yaml

scripts/solo_learn_dt/cifar10/barlow_diff.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="cifar_single"
+#SBATCH --qos=m2
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:4
+#SBATCH --ntasks-per-node=2
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=4:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/cifar10/ \
+    --config-name barlow_diff.yaml

scripts/solo_learn_dt/cifar10/barlow_icgan.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="cifar_single"
+#SBATCH --qos=m2
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:4
+#SBATCH --ntasks-per-node=2
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=4:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/cifar10/ \
+    --config-name barlow_icgan.yaml

scripts/solo_learn_dt/cifar10/byol.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="cifar_single"
+#SBATCH --qos=m2
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:4
+#SBATCH --ntasks-per-node=2
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=4:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/cifar10/ \
+    --config-name byol.yaml

scripts/solo_learn_dt/cifar10/byol_diff.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="cifar_single"
+#SBATCH --qos=m2
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:4
+#SBATCH --ntasks-per-node=2
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=4:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/cifar10/ \
+    --config-name byol_diff.yaml

scripts/solo_learn_dt/cifar10/byol_icgan.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="cifar_single"
+#SBATCH --qos=m2
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:4
+#SBATCH --ntasks-per-node=2
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=4:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/cifar10/ \
+    --config-name byol_icgan.yaml

scripts/solo_learn_dt/cifar10/moco.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="cifar_single"
+#SBATCH --qos=m2
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:4
+#SBATCH --ntasks-per-node=2
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=4:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/cifar10/ \
+    --config-name moco.yaml

scripts/solo_learn_dt/cifar10/moco_diff.slrm

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+#SBATCH --job-name="cifar_single"
+#SBATCH --qos=m2
+#SBATCH --nodes=1
+#SBATCH --gres=gpu:4
+#SBATCH --ntasks-per-node=2
+#SBATCH --cpus-per-task=8
+#SBATCH --mem=0
+#SBATCH --output=singlenode-eval-%j.out
+#SBATCH --error=singlenode-eval-%j.err
+#SBATCH --open-mode=append
+#SBATCH --wait-all-nodes=1
+#SBATCH --time=4:00:00
+
+# 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 CUDA_LAUNCH_BLOCKING=1
+
+export PYTHONPATH="."
+nvidia-smi
+
+
+torchrun --nproc-per-node=4 --nnodes=1 solo-learn/main_linear.py \
+    --config-path scripts/linear/cifar10/ \
+    --config-name moco_diff.yaml