Skip to content

Pull requests: huggingface/kernels-community

New pull request New
20 Open 1,014 Closed
20 Open 1,014 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

feat: cv-utils stable abi feature New functionality / capability
#1080 opened Aug 14, 2026 by drbh Collaborator Loading…
add conv1D kernel for XPU chore Version bumps, releases, misc maintenance
#1079 opened Aug 14, 2026 by kaixuanliu Contributor • Draft
mamba-ssm: dispatch Triton launchers on the tensor's device module chore Version bumps, releases, misc maintenance
#1078 opened Aug 13, 2026 by jiqing-feng Contributor • Draft
mamba-ssm: fix selective_state_update launch args when D or dt_bias is None chore Version bumps, releases, misc maintenance
#1077 opened Aug 13, 2026 by jiqing-feng Contributor • Draft
triton-kernels: make matmul_ogs_torch usable off CUDA and with default round_x chore Version bumps, releases, misc maintenance
#1076 opened Aug 13, 2026 by jiqing-feng Contributor • Draft
triton-kernels: fix Intel opt-flag heuristics copied from the NVIDIA path chore Version bumps, releases, misc maintenance
#1075 opened Aug 13, 2026 by jiqing-feng Contributor • Draft
fla: fix relative import depth in triton_ascend backend chore Version bumps, releases, misc maintenance
#1074 opened Aug 13, 2026 by jiqing-feng Contributor • Draft
megablocks: run native MXFP4 experts through the XPU grouped GEMM chore Version bumps, releases, misc maintenance
#1073 opened Aug 12, 2026 by jiqing-feng Contributor Loading…
1
build(deps): bump the actions group with 2 updates dependencies Pull requests that update a dependency file deps Dependency bumps (mostly dependabot) github_actions Pull requests that update GitHub Actions code
#1072 opened Aug 11, 2026 by dependabot Bot Loading…
activation: Add XPU backend to kernels-community/activation chore Version bumps, releases, misc maintenance
#1071 opened Aug 10, 2026 by jiqing-feng Contributor Loading…
1
Megablocks mxfp8 cri chore Version bumps, releases, misc maintenance
#1066 opened Aug 6, 2026 by jiqing-feng Contributor Loading…
flash-attn2: fix forward epilogue reduction for sycl-tla 0.9.2 chore Version bumps, releases, misc maintenance
#1065 opened Aug 6, 2026 by jiqing-feng Contributor Loading…
sgl-flash-attn3: switch to torch stable ABI chore Version bumps, releases, misc maintenance
#1054 opened Jul 31, 2026 by adarshxs Contributor Loading…
6 tasks
13
finegrained-kernels: 2D/Batched/Grouped GEMMs+Epilogues, SwapAB, Warp Spec, Offline/Inline Quant, all precisions and quantization fp32/fp16/bf16/fp8/mxfp8/mxfp4/nvfp4/weight-only 4-bit chore Version bumps, releases, misc maintenance
#1018 opened Jul 8, 2026 by IlyasMoutawwakil Member Loading…
6 tasks
54
flash-attn3: improve build speed, switch to get_kernel for tests build Build system: nix flakes, kernel-builder build, packaging cuda NVIDIA CUDA kernels performance Autotuning, speedups, build-speed improvements
#999 opened Jun 29, 2026 by danieldk Member Loading…
3
vllm-quantization: import in kernels-community again cuda NVIDIA CUDA kernels feature New functionality / capability new-kernel A brand-new kernel package is added rocm AMD / AITER kernels
#991 opened Jun 23, 2026 by danieldk Member Loading…
tokenspeed-attention: add TokenSpeed MHA/MLA Triton kernels feature New functionality / capability new-kernel A brand-new kernel package is added triton Vendor-neutral Triton kernels upstream-sync Syncing a kernel with its upstream source
#990 opened Jun 23, 2026 by adarshxs Contributor • Draft
2
ci: sign metadata for Windows builds ci Workflows, GitHub Actions, automation
#984 opened Jun 22, 2026 by danieldk Member Loading…
2
tdt-loss: add TDT loss kernel cuda NVIDIA CUDA kernels feature New functionality / capability new-kernel A brand-new kernel package is added
#882 opened May 19, 2026 by ebezzam Loading…
4 tasks
1
make ScatterMoe more easily patchable for HF MoE modeling cuda NVIDIA CUDA kernels feature New functionality / capability new-layer Adds or extends a LayerRepo
#82 opened Dec 5, 2025 by winglian Loading…
12
ProTip! Adding no:label will show everything without a label.