Kernels

Commit History

  1. Update tag⁠

    danieldk HF Staff
    Verified signature
    3083319

  1. Add built binary [skip-build]

    github-actions[bot]
    f2471cd

  1. fix(rms_norm.py): add assertion for input gradients to handle unsupported cases in backward pass

    f19f8f4
  2. feat: support sequence parallel with fused_add_rms_norm

    151bb5a
  3. refactor(activation): change fused_add_rms_norm and fused_add_rms_norm_backward to out-place operations

    7e4334d
  4. refactor(rms_norm): move RMS normalization logic to a new module for better organization and maintainability

    66b3c5e

  1. feat: support sequence parallel with rms_norm

    06d6367
  2. feat: add assert is_contiguous

    a2a2501
  3. feat: make rms_norm as out-place

    9d0a235

  1. feat(workflow): add Slack notifications for build start, success, and failure [skip-build]

    ab05e35
  2. Revert "fix typo in readme (#7)" (#8)⁠

    TaehyunKim
    Unverified signature
    ddd119c
  3. fix typo in readme (#7)⁠

    TaehyunKim +1: github-actions[bot]
    Unverified signature
    2d926c3

  1. chore: add build action [skip-build]

    5dde6fa

  1. Fix fused add rms norm (#4)⁠

    TaehyunKim +1: TaehyunKimMotif
    Unverified signature
    a1e5ca8

  1. Add fusion (#3)⁠

    TaehyunKim +1: TaehyunKimMotif
    Unverified signature
    e5e2eeb

  1. chore: add license

    e677f62
  2. chore: add push-to-hf workflow

    269223c

  1. Merge pull request #1 from MotifTechnologies/add_action⁠

    TaehyunKim
    Unverified signature
    9c49e08
  2. add more exclude files

    f9f3398
  3. add precommit acttion

    294dbc1

  1. Optimize kernel (#2)⁠

    Verified signature
    97825b8

  1. torch 2.8 support (#1)⁠

    Verified signature
    43629b7

  1. feat: add cuda build

    cf68df1

  1. feat: support reset_parameters()

    605f22e
  2. feat(rms-norm): Impl fused RMSNorm

    f3b99fb

  1. feat(poly-norm): add perf test

    d14fd4d

  1. fix(poly-norm): calc param grad explicitly

    704692b
  2. fix(poly-norm): fix bug in reduce sum

    883cc1c
  3. refactor(poly-norm): fix indentation

    32c2bde

  1. refactor(poly-norm): use const for immutable args

    e85ecc9
  2. chore(poly-norm): remove unnecessary file

    552d415
  3. chore: use latest build image and misc

    f5a7d38

  1. chore(poly-norm): update README and build artifacts

    f72121c
  2. feat(poly-norm): add default value for eps argument

    afd2a56
  3. chore(poly-norm): add ROCm build artifacts

    4b70498
  4. feat(poly-norm): Add PolyNorm

    44e9845
  5. initial commit

    Verified signature
    7a7d761