An open API service for producing an overview of a list of open source projects.

https://github.com/NVIDIA/cudnn-frontend

attention blackwell cuda cuda-kernels cuda-toolkit deep-learning flash-attention fp8 gemm gpu grouped-gemm hopper mixture-of-experts moe mxfp8 normalization nvfp4 nvidia sdpa transformer

Score: 26.847615647679397

Last synced: 17 minutes ago
JSON representation

Repository metadata:

cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.


Owner metadata:


GitHub Events

Total
Last Year

Committers metadata

Last synced: 1 day ago

Total Commits: 1,594
Total Committers: 115
Avg Commits per committer: 13.861
Development Distribution Score (DDS): 0.721

Commits in past year: 1,505
Committers in past year: 108
Avg Commits per committer in past year: 13.935
Development Distribution Score (DDS) in past year: 0.705

Name Email Commits
Yang Xu 3****c 444
Anerudhan Gopal a****l@n****m 183
RomanAnders90 2****0 134
Vedaanta Agarwalla 1****a 97
0z5a d****u@s****e 85
Hwanseo Choi h****c@n****m 70
Yanqin Zhai y****i@g****m 47
Yihua Wei y****w@n****m 42
Emil Gilliam e****m@n****m 39
jiayus-nvidia j****s@n****m 39
yanzhuo607 y****c@n****m 31
Chaofan Yu 1****n 31
Brandon Zhang 3****g 31
Josh Park 8****k 26
BARRET 4****s 24
Hiki a****2@g****m 23
Ye Liu y****u@n****m 13
tongke 1****6 11
zky k****u@z****i 11
pmdavies-nv p****s@n****m 8
Jane (Jiancheng) Liu l****e@n****m 8
msalasooNV 2****V 8
Butterfingrz 1****z 7
adshen a****n@n****m 7
meinie m****s@g****m 7
Sun 2****0@q****m 6
miyan 1****8@q****m 6
Siddhartha Raman Sundara Raman s****n@n****m 5
Ziang Li z****i@u****u 5
Zixin Huang z****h@n****m 5
and 85 more...

Issue and Pull Request metadata

Last synced: about 1 hour ago

Total issues: 232
Total pull requests: 592
Average time to close issues: 3 months
Average time to close pull requests: 5 days
Total issue authors: 103
Total pull request authors: 101
Average comments per issue: 1.48
Average comments per pull request: 4.39
Merged pull request: 349
Bot issues: 0
Bot pull requests: 0

Past year issues: 137
Past year pull requests: 484
Past year average time to close issues: 8 days
Past year average time to close pull requests: 4 days
Past year issue authors: 36
Past year pull request authors: 81
Past year average comments per issue: 1.11
Past year average comments per pull request: 5.26
Past year merged pull request: 262
Past year bot issues: 0
Past year bot pull requests: 0

More stats: https://issues.ecosyste.ms/repositories/lookup?url=https://github.com/NVIDIA/cudnn-frontend

Top Issue Authors

  • liujane-dev (54)
  • vedaanta (35)
  • Anerudhan (10)
  • vedantroy (7)
  • YangXu1990uiuc (4)
  • zhewenhu (4)
  • akineeic (3)
  • zcbenz (3)
  • swimvtec (3)
  • Skylion007 (3)
  • iskunk (3)
  • EnricoDeg (2)
  • gritukan (2)
  • JackRao123 (2)
  • xcwang1999 (2)

Top Pull Request Authors

  • Anerudhan (130)
  • YangXu1990uiuc (96)
  • vedaanta (51)
  • hwanseoc (26)
  • brandonfzhang (18)
  • jiayus-nvidia (16)
  • yanqinz2 (16)
  • egilliam-nv (15)
  • RomanAnders90 (13)
  • zkyue (11)
  • jhjpark (11)
  • Adnios (10)
  • yeliu-oss (9)
  • yanzhuo607 (8)
  • 0z5a (7)

Top Issue Labels

  • cat-bug (65)
  • orig-qa (55)
  • mod-frontend (31)
  • orig-nv-eng (28)
  • cat-feature (16)
  • area:global_attention (16)
  • cat-ci (15)
  • cat-enhancements (14)
  • mod-cutedsl (13)
  • mod-frost (10)
  • mod-backend (8)
  • open-todo (7)
  • orig-external (6)
  • open-unassigned (5)
  • cudnn_frontend-1.17.0 (3)
  • cat-perf-bug (2)
  • open-verify-to-close (2)
  • area:frost (2)
  • cat-doc (2)
  • cudnn-frontend-1.5.0 (2)
  • cat-infra (1)
  • enhancement (1)
  • roadmap (1)
  • mod-infra (1)
  • cat-bugfix (1)
  • cat-question (1)

Top Pull Request Labels

  • orig-nv-eng (286)
  • mod-cutedsl (110)
  • cat-enhancements (107)
  • mod-frost (98)
  • cat-feature (85)
  • mod-frontend (84)
  • area:frost (80)
  • area:global_attention (67)
  • cat-bugfix (46)
  • orig-external (41)
  • cat-perf-bug (31)
  • cat-bug (29)
  • area:sparse_attention (25)
  • cat-ci (23)
  • cat-cleanup (19)
  • mod-infra (15)
  • cat-infra (15)
  • op: DSA (15)
  • area:linear_attention (13)
  • mod-backend (12)
  • cat-doc (9)
  • stale (7)
  • orig-qa (7)
  • cat-routine-update (7)
  • area:packaging (1)
  • routine_maintenance (1)
  • open-todo (1)
  • closed-fixed (1)
  • closed-stale (1)

Package metadata

proxy.golang.org: github.com/nvidia/cudnn-frontend

  • Homepage:
  • Documentation: https://pkg.go.dev/github.com/nvidia/cudnn-frontend#section-documentation
  • Licenses: apache-2.0
  • Latest release: v1.32.1 (published 5 days ago)
  • Last Synced: 2026-11-05T22:41:35.928Z (1 day ago)
  • Versions: 68
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Rankings:
    • Stargazers count: 3.531%
    • Forks count: 3.534%
    • Average: 6.861%
    • Dependent packages count: 9.576%
    • Dependent repos count: 10.802%
proxy.golang.org: github.com/NVIDIA/cudnn-frontend

  • Homepage:
  • Documentation: https://pkg.go.dev/github.com/NVIDIA/cudnn-frontend#section-documentation
  • Licenses: apache-2.0
  • Latest release: v1.32.1 (published 8 days ago)
  • Last Synced: 2026-11-05T22:41:35.407Z (1 day ago)
  • Versions: 68
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Rankings:
    • Forks count: 3.562%
    • Stargazers count: 3.576%
    • Average: 6.879%
    • Dependent packages count: 9.576%
    • Dependent repos count: 10.802%
pypi.org: nvidia-cudnn-frontend

NVIDIA cuDNN Frontend — Python and C++ Graph API with SOTA attention (SDPA / Flash Attention), MoE grouped GEMM fusions, and FP8/MXFP8 kernels for Hopper, Blackwell, and Rubin GPUs.

  • Homepage: https://github.com/NVIDIA/cudnn-frontend
  • Documentation: https://docs.nvidia.com/deeplearning/cudnn/frontend/latest/
  • Licenses: Apache-2.0 AND MIT
  • Latest release: 1.32.2 (published 2 days ago)
  • Last Synced: 2026-11-05T22:41:35.658Z (1 day ago)
  • Versions: 40
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Downloads: 3,232,266 Last month
  • Rankings:
    • Stargazers count: 4.682%
    • Forks count: 6.853%
    • Dependent packages count: 9.454%
    • Average: 20.84%
    • Dependent repos count: 62.372%
  • Maintainers (1)
anaconda.org: python-cudnn-frontend

cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels — scaled dot-product attention (SDPA / Flash Attention), grouped GEMM fusions for Mixture-of-Experts (MoE) training, fused normalization + activation, and more.

  • Homepage: https://github.com/NVIDIA/cudnn-frontend
  • Licenses: MIT
  • Latest release: 1.18.0 (published 4 months ago)
  • Last Synced: 2026-07-15T22:24:13.376Z (4 months ago)
  • Versions: 1
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Downloads: 100 Total
  • Rankings:
    • Dependent packages count: 34.375%
    • Dependent repos count: 37.481%
    • Average: 43.041%
    • Downloads: 57.267%
nixpkgs-unstable: python313Packages.nvidia-cudnn-frontend

Python and C++ Graph API with SOTA attention (SDPA / Flash Attention), MoE grouped GEMM fusions, and FP8/MXFP8 kernels for Hopper and Blackwell GPUs

nixpkgs-unstable: python314Packages.nvidia-cudnn-frontend

Python and C++ Graph API with SOTA attention (SDPA / Flash Attention), MoE grouped GEMM fusions, and FP8/MXFP8 kernels for Hopper and Blackwell GPUs

debian-13: nvidia-cudnn-frontend

  • Homepage: https://github.com/NVIDIA/cudnn-frontend
  • Documentation: https://packages.debian.org/trixie/nvidia-cudnn-frontend
  • Licenses: apache-2.0
  • Latest release: 1.8.0+ds-1 (published 9 months ago)
  • Last Synced: 2026-03-12T17:11:29.627Z (8 months ago)
  • Versions: 1
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Rankings:
    • Dependent repos count: 0.0%
    • Dependent packages count: 0.0%
    • Average: 100%
ubuntu-24.10: nvidia-cudnn-frontend

  • Homepage: https://github.com/NVIDIA/cudnn-frontend
  • Licenses: apache-2.0
  • Latest release: 0.9.2+ds-1 (published 9 months ago)
  • Last Synced: 2026-03-13T19:19:00.015Z (8 months ago)
  • Versions: 1
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Rankings:
    • Dependent repos count: 0.0%
    • Dependent packages count: 0.0%
    • Average: 100%
ubuntu-23.10: nvidia-cudnn-frontend

  • Homepage: https://github.com/NVIDIA/cudnn-frontend
  • Licenses: apache-2.0
  • Latest release: 0.8+ds-1 (published 9 months ago)
  • Last Synced: 2026-08-20T02:31:08.173Z (3 months ago)
  • Versions: 1
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Rankings:
    • Dependent repos count: 0.0%
    • Dependent packages count: 0.0%
    • Average: 100%
gentoo-portage: sci-ml/cudnn-frontend

A c++ wrapper for the cudnn backend API

  • Homepage: https://github.com/NVIDIA/cudnn-frontend
  • Documentation: https://packages.gentoo.org/packages/sci-ml/cudnn-frontend
  • Licenses: MIT
  • Latest release: 1.16.1 (published 7 months ago)
  • Last Synced: 2026-05-27T04:28:41.630Z (5 months ago)
  • Versions: 5
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Rankings:
    • Dependent repos count: 0.0%
    • Dependent packages count: 0.0%
    • Average: 100%
nixpkgs-unstable: cudaPackages.cudnn-frontend

Python and C++ Graph API with SOTA attention (SDPA / Flash Attention), MoE grouped GEMM fusions, and FP8/MXFP8 kernels for Hopper and Blackwell GPUs

ubuntu-24.04: nvidia-cudnn-frontend

  • Homepage: https://github.com/NVIDIA/cudnn-frontend
  • Licenses: apache-2.0
  • Latest release: 0.9.2+ds-1 (published 9 months ago)
  • Last Synced: 2026-08-08T10:10:49.908Z (3 months ago)
  • Versions: 1
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Rankings:
    • Dependent repos count: 0.0%
    • Dependent packages count: 0.0%
    • Average: 100%
registry.bazel.build: cudnn_frontend

  • Homepage: https://github.com/NVIDIA/cudnn-frontend
  • Documentation: https://registry.bazel.build/docs/cudnn_frontend
  • Licenses: apache-2.0
  • Latest release: 1.12.0 (published about 1 month ago)
  • Last Synced: 2026-11-05T22:41:33.305Z (1 day ago)
  • Versions: 1
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Rankings:
    • Dependent repos count: 0.0%
    • Dependent packages count: 0.0%
    • Average: 100%
  • Maintainers (1)

Dependencies

benchmark/linear_attention/Dockerfile docker
  • nvcr.io/nvidia/pytorch 26.03-py3 build
pyproject.toml pypi
requirements.txt pypi
  • black ==26.3.1
  • clang-format ==21.1.6
  • jupyter *
  • looseversion *
  • numpy <2.0.0
  • pybind11 >=2.13,<3
  • pytest *
  • pytest-xdist *
setup.py pypi
tools/cudnn_repro/pyproject.toml pypi
benchmark/norms/Dockerfile docker
  • nvcr.io/nvidia/pytorch 25.12-py3 build
.github/workflows/stale.yml actions
  • actions/stale v9 composite
benchmark/e2e/Qwen-Image/requirements.txt pypi
.github/workflows/pr-merge-requirements.yml actions
benchmark/attention_training/Dockerfile docker
  • nvcr.io/nvidia/pytorch 26.03-py3 build
.github/workflows/pre-commit.yml actions
  • actions/cache v4 composite
  • actions/checkout v4 composite
  • actions/setup-python v5 composite
test/python/block_sparse_attention/jax/Dockerfile docker
  • nvcr.io/nvidia/jax 26.07-py3@sha256 build
.github/workflows/docs-fern-publish.yml actions
  • actions/checkout v4 composite
  • actions/setup-node v4 composite
benchmark/e2e/Wan2.2-VAE/requirements.txt pypi
  • diffusers ==0.40.0
.github/workflows/docs-fern.yml actions
  • actions/checkout v4 composite
  • actions/setup-node v4 composite