Skip to content

Actions: bitsandbytes-foundation/bitsandbytes

Actions

Python package

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
python-package.yml
1,112 workflow runs
1,112 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
[ROCm] Enable blocksize 32 4-bit quantization and GEMV kernels on AMD… Python package #2454: Commit 713a3b8 pushed by matthewdouglas
9m 30s main
main
9m 30s
cuda_specs.py: fix version tag string for ROCm/CUDA library lookup (#… Python package #2453: Commit f1dcf42 pushed by matthewdouglas
2m 6s main
main
2m 6s
Add CUDA graph support and consolidate dense+MoE GEMM into single TU Python package #2452: Commit 5312321 pushed by TimDettmers
7m 22s feature/moe-batched-sm120
feature/moe-batched-sm120
7m 22s
[ROCm] Enable blocksize 32 4-bit quantization and GEMV kernels on AMD CDNA Python package #2451: Pull request #1887 synchronize by sstamenk
10m 30s sstamenk:fix/rocm-cdna-blocksize-32-64
sstamenk:fix/rocm-cdna-blocksize-32-64
10m 30s
Create SECURITY.md (#1892) Python package #2450: Commit 27cfc90 pushed by matthewdouglas
10m 24s main
main
10m 24s
Create SECURITY.md Python package #2449: Commit a650311 pushed by matthewdouglas
10m 48s create-security-policy
create-security-policy
10m 48s
Add batched NVFP4 MoE GEMM for SM_100 (B200) with benchmark Python package #2448: Commit c872065 pushed by TimDettmers
7m 41s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
7m 41s
Use worst-case buffer sizing for grouped GEMM cache Python package #2447: Commit f8aa403 pushed by TimDettmers
8m 16s feature/moe-batched-sm120
feature/moe-batched-sm120
8m 16s
Use worst-case buffer sizing for grouped GEMM cache Python package #2446: Commit f8aa403 pushed by TimDettmers
8m 20s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
8m 20s
Remove redundant SFB un-swizzle/re-swizzle and return BF16 directly Python package #2445: Commit 288d1f0 pushed by TimDettmers
7m 3s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
7m 3s
Add SM_100 (Blackwell B200) CUTLASS grouped GEMM with fused C dispatch Python package #2444: Commit 3c4acbd pushed by TimDettmers
8m 6s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
8m 6s
Rewrite grouped MoE GEMM to use swizzled (block-scaled) scales Python package #2443: Commit 33116da pushed by TimDettmers
8m 54s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
8m 54s
Use swizzled scales exclusively for all dense NVFP4 GEMM paths Python package #2442: Commit e6c5900 pushed by TimDettmers
54m 57s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
54m 57s
Rewrite hand-written NVFP4 GEMM to use swizzled (block-scaled) scales Python package #2441: Commit b52ab38 pushed by TimDettmers
3m 13s feature/qutlass-nvfp4-gemm
feature/qutlass-nvfp4-gemm
3m 13s
cuda_specs.py: fix version tag string for ROCm/CUDA library lookup Python package #2440: Pull request #1889 opened by eliasmagn
11m 6s eliasmagn:main
eliasmagn:main
11m 6s
[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
ProTip! You can narrow down the results and go further in time using created:<2026-03-05 or the other filters available.