Commits

Branch selector

 master

User selector

All users

All time

Commit History

Commits on Oct 26, 2025

• modify-readme-2

  Source-Roc

  committed
  093821b

  Copy full SHA for 093821b

• modify tensor.type() to tensor.options()

  Source-Roc

  committed
  9821ea5

  Copy full SHA for 9821ea5

• Finish cleaning up usage of VERSION_GE_1_1/1_3/1_5 macros

  Source-Roc

  committed
  82c4224

  Copy full SHA for 82c4224

• modify-readme

  Source-Roc

  committed
  5bb565e

  Copy full SHA for 5bb565e

Commits on Oct 3, 2025

• fix compile group norm [v1/v2] (#1927)

  Show description for 5c8860f
  AlongWY

  authored
  5c8860f

  Copy full SHA for 5c8860f

Commits on Sep 29, 2025

• [contrib-sparsity] use `os.getenv` for simplicity (#1933)

  Aidyn-A

  authored
  eb1e7d9

  Copy full SHA for eb1e7d9

Commits on Sep 25, 2025

• Update group norm license. (#1935)

  alpha0422

  authored
  4fd4f30

  Copy full SHA for 4fd4f30

Commits on Sep 5, 2025

• Remove pickle, minor variable rename (#1924)

  Show description for 214c6b3
  jpool-nv

  authored
  214c6b3

  Copy full SHA for 214c6b3

• Use Pytorch's AMP for amp test (#1929)

  tinglvv

  authored
  c54dbd3

  Copy full SHA for c54dbd3

Commits on Aug 21, 2025

• Remove version dependent macros & extension specific arch flags except fmha (#1918)

  crcrpar

  authored
  26bba57

  Copy full SHA for 26bba57

Commits on Aug 4, 2025

• [contrib/torchsched] Refactor cross-stream dependency handling (#1922)

  Show description for 4bdecd0
  

  lirundong

  and

  alpha0422

  authored
  4bdecd0

  Copy full SHA for 4bdecd0

Commits on Jul 30, 2025

• Update setup.py for 13.0 (#1921)

  Show description for 446a6c1
  eqy

  authored
  446a6c1

  Copy full SHA for 446a6c1

• revert previous commit

  eqy

  authored
  0654ec3

  Copy full SHA for 0654ec3

• Update setup.py

  eqy

  authored
  50a671c

  Copy full SHA for 50a671c

Commits on Jul 29, 2025

• Handle temporary arg name divergence in the `torch.cuda.MemPool` (#1920)

  Show description for bf903a2
  youngeunkwon0405

  authored
  bf903a2

  Copy full SHA for bf903a2

Commits on Jul 22, 2025

• Allow env vars to specify custom extensions to build (#1910)

  Show description for bfb500c
  crcrpar

  authored
  bfb500c

  Copy full SHA for bfb500c

Commits on Jul 12, 2025

• [contrib/nccl_allocator] Add an interface for the NCCL symmetric registration (#1915)

  Show description for 2eda0ac
  

  youngeunkwon0405

  and

  crcrpar

  authored
  2eda0ac

  Copy full SHA for 2eda0ac

Commits on Jun 26, 2025

• replace `apex.parallel.flat_dist_call` in test with coalescing manager (#1913)

  Show description for d8200a1
  crcrpar

  authored
  d8200a1

  Copy full SHA for d8200a1

Commits on Jun 25, 2025

• fmha: add sm_100 and sm_120 (#1890)

  Aidyn-A

  authored
  a255aa6

  Copy full SHA for a255aa6

• Update test_fast_layer_norm.py (#1911)

  eqy

  authored
  6445b06

  Copy full SHA for 6445b06

Commits on May 15, 2025

• [Contrib][Openfold Triton] Use `json` instead of `pickle` (#1900)

  Show description for e74a67b
  

  Aidyn-A

  and

  crcrpar

  authored
  e74a67b

  Copy full SHA for e74a67b

Commits on May 13, 2025

• Deprecate `apex.transformer` & `apex.contrib.gpu_direct_storage` (#1901)

  Show description for 8ad740d
  crcrpar

  authored
  8ad740d

  Copy full SHA for 8ad740d

Commits on May 7, 2025

• remove old features: amp, ddp, and rnn (#1896)

  Show description for cb65b8e
  crcrpar

  authored
  cb65b8e

  Copy full SHA for cb65b8e

Commits on Apr 11, 2025

• Graph Compiler For Multi-stream (#1886)

  Show description for e13873d
  

  alpha0422

  and

  lirundong

  authored
  e13873d

  Copy full SHA for e13873d

Commits on Apr 10, 2025

• Pipeline `reduce-scatter` and `all-reduce`. (#1895)

  alpha0422

  authored
  b9d758c

  Copy full SHA for b9d758c

• Fix GroupNorm test: properly check graph breaks (#1894)

  yuantailing

  authored
  badf311

  Copy full SHA for badf311

Commits on Apr 9, 2025

• Fix `DistributedFusedAdam` for grad dtype != param dtype (#1893)

  yuantailing

  authored
  2eb64da

  Copy full SHA for 2eb64da

Commits on Apr 4, 2025

• [contrib] torch.compile pass for replacing layer norm with cudnn version. (#1892)

  alpha0422

  authored
  379937e

  Copy full SHA for 379937e

Commits on Mar 26, 2025

• [contrib] Add group_norm_v2 (#1887)

  Show description for c02c6c8
  yuantailing

  authored
  c02c6c8

  Copy full SHA for c02c6c8

Commits on Mar 18, 2025

• Support CUDA12.8 in `setup.py` (#1889)

  Show description for 8c23de0
  

  4 peopleauthored
  8c23de0

  Copy full SHA for 8c23de0

Commits on Mar 14, 2025

• [contrib] Improve `apex.contrib.focal_loss` for B200 (#1888)

  Show description for 9c50239
  

  crcrpar

  and

  minitu

  authored
  9c50239

  Copy full SHA for 9c50239

Commits on Feb 26, 2025

• [nccl_allocator] Adds helper API to create pool (#1877)

  syed-ahmed

  authored
  312acb4

  Copy full SHA for 312acb4

• adjust threshold for `test_encdec_multihead_attn.py` (#1884)

  Fuzzkatt

  authored
  0024663

  Copy full SHA for 0024663

Commits on Feb 25, 2025

• GroupNorm: add 384-channel support (#1883)

  Show description for bb9dccd
  

  yuantailing

  and

  crcrpar

  authored
  bb9dccd

  Copy full SHA for bb9dccd

• Add option to build extensions in parallel (#1882)

  Show description for c9e6f05
  yuantailing

  authored
  c9e6f05

  Copy full SHA for c9e6f05