Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Nov 28, 2025

• Enable Ruff pre-commit hooks (#1957)

  Show description for 16fadfe
  

  3 peopleauthored
  16fadfe

  Copy full SHA for 16fadfe

Commits on Oct 29, 2025

• Fix condition to fallback in MixedDtype RMS|Layer Norm to fix export test failures (#1940)

  Show description for 718fcf2
  crcrpar

  authored
  718fcf2

  Copy full SHA for 718fcf2

Commits on Dec 14, 2024

• Traceable LayerNorm (#1864)

  yanboliang

  authored
  73375b3

  Copy full SHA for 73375b3

Commits on Dec 9, 2024

• Traceable RMSNorm part 2: non affine fused rms norm (#1862)

  Show description for 2863aa0
  yanboliang

  authored
  2863aa0

  Copy full SHA for 2863aa0

Commits on Nov 28, 2024

• Traceable RMSNorm (#1861)

  yanboliang

  authored
  ac8214e

  Copy full SHA for ac8214e

Commits on Nov 1, 2024

• manual impl when compiling (#1854)

  alexdremov

  authored
  e5e8b89

  Copy full SHA for e5e8b89

Commits on Jun 29, 2024

• deprecate uses of torch.cuda.amp (#1813)

  Show description for f8e60c4
  Fuzzkatt

  authored
  f8e60c4

  Copy full SHA for f8e60c4

Commits on Jan 1, 2024

• Make fused layer norm functions backward-compatible (#1760)

  Show description for c07a4cf
  timmoon10

  authored
  c07a4cf

  Copy full SHA for c07a4cf

Commits on Sep 28, 2023

• Massively reduce LayerNorm/RMSNorm GPU memory usage in modern networks by tricking torch autograd (#1715)

  Show description for 6ff4548
  RuiWang1998

  authored
  6ff4548

  Copy full SHA for 6ff4548

Commits on Apr 27, 2023

• Using regular layernorm for ONNX and TS export (#1634)

  Show description for 85e9edd
  

  borisfom

  and

  crcrpar

  authored
  85e9edd

  Copy full SHA for 85e9edd

Commits on Feb 13, 2023

• replace torch.Tensor with torch.empty (#1578)

  Show description for ba027dd
  NouamaneTazi

  authored
  ba027dd

  Copy full SHA for ba027dd

Commits on Sep 14, 2022

• Update test_fused_layer_norm.py (#1485)

  Show description for 136a13c
  Fuzzkatt

  authored
  136a13c

  Copy full SHA for 136a13c

Commits on Feb 11, 2022

• [FusedRMSNorm doc] document where epsilon is added (#1295)

  Show description for c8c00ef
  stas00

  authored
  c8c00ef

  Copy full SHA for c8c00ef

Commits on Feb 7, 2022

• fix and generate docs for FusedRMSNorm (#1285)

  eqy

  authored
  a786ca0

  Copy full SHA for a786ca0

Commits on Feb 4, 2022

• FusedRMSNorm/"T5LayerNorm" based on FusedLayerNorm (#1274)

  Show description for 684c473
  

  eqy

  and

  crcrpar

  authored
  684c473

  Copy full SHA for 684c473

Commits on Oct 2, 2021

• transformer utils (#1181)

  Show description for 365fdc1
  

  5 peopleauthored
  365fdc1

  Copy full SHA for 365fdc1

Commits on Jul 31, 2019

• use ctx.eps instead of eps

  

  ngimel

  authored
  0dbf6c2

  Copy full SHA for 0dbf6c2

Commits on Jul 20, 2019

• Update FusedLayerNorm for new function API

  

  Myle Ott

  committed
  d1626cc

  Copy full SHA for d1626cc

Commits on Apr 10, 2019

• quick fix: make FusedLayerNorm compatible with cpu

  LamDang

  committed
  d130ec1

  Copy full SHA for d130ec1

Commits on Mar 12, 2019

• Forward/backward compatibility around pytorch 3aeb78, to fix #191

  definitelynotmcarilli

  committed
  42180bd

  Copy full SHA for 42180bd

Commits on Feb 3, 2019

• Lazy imports to reduce error spam

  definitelynotmcarilli

  committed
  48299b0

  Copy full SHA for 48299b0

Commits on Nov 28, 2018

• minor latex touchup

  definitelynotmcarilli

  committed
  b436213

  Copy full SHA for b436213

Commits on Oct 31, 2018

• [WIP] Fused layer norm cuda (#69)

  Show description for 1b9b65c
  

  thorjohnsen

  authored and

  mcarilli

  committed
  1b9b65c

  Copy full SHA for 1b9b65c

Loading