Skip to content

Actions: bitsandbytes-foundation/bitsandbytes

Actions

Python package

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
python-package.yml
1,179 workflow runs
1,179 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

cuda_specs.py: fix version tag string for ROCm/CUDA library lookup Python package #2440: Pull request #1889 opened by eliasmagn
Action required eliasmagn:main
eliasmagn:main
Action required
[ROCm] Enable blocksize 32 4-bit quantization and GEMV kernels on AMD CDNA Python package #2439: Pull request #1887 synchronize by sstamenk
21m 7s sstamenk:fix/rocm-cdna-blocksize-32-64
sstamenk:fix/rocm-cdna-blocksize-32-64
21m 7s
Enable more targets for ROCm (#1886) Python package #2438: Commit 3322a64 pushed by matthewdouglas
16m 33s main
main
16m 33s
[ROCm] Replace compile-time warp size with runtime query in host code… Python package #2437: Commit 373f23b pushed by matthewdouglas
11m 15s main
main
11m 15s
[ROCm] Replace compile-time warp size with runtime query in host code Python package #2436: Pull request #1885 synchronize by sstamenk
8m 32s sstamenk:fix/rocm-runtime-warp-size
sstamenk:fix/rocm-runtime-warp-size
8m 32s
[ROCm] Replace compile-time warp size with runtime query in host code Python package #2435: Pull request #1885 synchronize by sstamenk
3m 30s sstamenk:fix/rocm-runtime-warp-size
sstamenk:fix/rocm-runtime-warp-size
3m 30s
[ROCm] Rework ROCm build to use ROCm version instead of HIP version Python package #2434: Pull request #1888 synchronize by sstamenk
11m 6s sstamenk:fix/rocm-build-rework
sstamenk:fix/rocm-build-rework
11m 6s
Remove dead cgemm_nvfp4_grouped FP32 entry point Python package #2433: Commit 4dbc169 pushed by TimDettmers
10m 39s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
10m 39s
Remove dead CUDA entry points: cgemm_nvfp4, cgemm_nvfp4_splitk, cgemm… Python package #2432: Commit 55bf8f3 pushed by TimDettmers
1m 36s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
1m 36s
Remove dead _gemm_nvfp4_hw_raw and _gemm_nvfp4_hw_splitk_raw wrappers Python package #2431: Commit f455bb8 pushed by TimDettmers
3m 29s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
3m 29s
Add full Python integration for grouped NVFP4 GEMM Python package #2430: Commit 42319cf pushed by TimDettmers
13m 26s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
13m 26s
Remove dead hand-written NVFP4 quantize/Hadamard kernels and simple GEMM Python package #2429: Commit fe2f52e pushed by TimDettmers
4m 42s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
4m 42s
Vectorize activation loads in VQ scalar GEMV, update to GLM-4.7 shapes Python package #2428: Commit ab79ad2 pushed by TimDettmers
9m 34s feature/kbit-gemv-v8
feature/kbit-gemv-v8
9m 34s
[ROCm] Rework ROCm build to use ROCm version instead of HIP version Python package #2427: Pull request #1888 synchronize by sstamenk
8m 26s sstamenk:fix/rocm-build-rework
sstamenk:fix/rocm-build-rework
8m 26s
[ROCm] Enable more targets Python package #2426: Pull request #1886 synchronize by sstamenk
10m 3s sstamenk:fix/rocm-enable-more-targets
sstamenk:fix/rocm-enable-more-targets
10m 3s
[ROCm] Rework ROCm build to use ROCm version instead of HIP version Python package #2425: Pull request #1888 synchronize by sstamenk
Action required sstamenk:fix/rocm-build-rework
sstamenk:fix/rocm-build-rework
Action required
[ROCm] Replace compile-time warp size with runtime query in host code Python package #2424: Pull request #1885 synchronize by sstamenk
8m 58s sstamenk:fix/rocm-runtime-warp-size
sstamenk:fix/rocm-runtime-warp-size
8m 58s
[ROCm] Rework ROCm build to use ROCm version instead of HIP version Python package #2423: Pull request #1888 synchronize by sstamenk
Action required sstamenk:fix/rocm-build-rework
sstamenk:fix/rocm-build-rework
Action required
[ROCm] Rework ROCm build to use ROCm version instead of HIP version Python package #2422: Pull request #1888 synchronize by sstamenk
Action required sstamenk:fix/rocm-build-rework
sstamenk:fix/rocm-build-rework
Action required
[ROCm] Rework ROCm build to use ROCm version instead of HIP version Python package #2421: Pull request #1888 synchronize by sstamenk
Action required sstamenk:fix/rocm-build-rework
sstamenk:fix/rocm-build-rework
Action required
[ROCm] Rework ROCm build to use ROCm version instead of HIP version Python package #2420: Pull request #1888 opened by sstamenk
Action required sstamenk:fix/rocm-build-rework
sstamenk:fix/rocm-build-rework
Action required
Add VQ grouped scalar GEMV kernel for MoE inference (M=1-4) Python package #2419: Commit 5f0bc6d pushed by TimDettmers
1h 5m 9s feature/moe-vq-kernel
feature/moe-vq-kernel
1h 5m 9s
feat: Add grouped NVFP4 GEMM kernel for MoE inference Python package #2418: Commit d90b8d3 pushed by TimDettmers
13m 14s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
13m 14s
Add VQ grouped GEMM benchmark Python package #2415: Commit ab66ce7 pushed by TimDettmers
9m 32s feature/moe-vq-kernel
feature/moe-vq-kernel
9m 32s
[ROCm] Enable blocksize 32 4-bit quantization and GEMV kernels on AMD CDNA Python package #2414: Pull request #1887 synchronize by sstamenk
1h 48m 39s sstamenk:fix/rocm-cdna-blocksize-32-64
sstamenk:fix/rocm-cdna-blocksize-32-64
1h 48m 39s