Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Nov 27, 2025

• Add pre-commit and GitHub Actions workflow for it (#1949)

  Show description for 1f20398
  crcrpar

  authored
  1f20398

  Copy full SHA for 1f20398

Commits on Nov 8, 2025

• Remove `compat.h` (#1942)

  Show description for e900daa
  

  crcrpar

  and

  Copilot

  authored
  e900daa

  Copy full SHA for e900daa

Commits on May 22, 2020

• more fixes on dtypes

  kexinyu

  committed
  cf918ac

  Copy full SHA for cf918ac

• use pointer

  kexinyu

  committed
  06a83ce

  Copy full SHA for 06a83ce

• .data<...>()

  kexinyu

  committed
  3a727a0

  Copy full SHA for 3a727a0

• at::Tensor::data_ptr()

  kexinyu

  committed
  2c3f3d9

  Copy full SHA for 2c3f3d9

• fix dtype

  kexinyu

  committed
  abc991d

  Copy full SHA for abc991d

Commits on May 21, 2020

• make fused LAMB async

  kexinyu

  committed
  f54cc1c

  Copy full SHA for f54cc1c

Commits on May 1, 2020

• Merge branch 'master' of https://github.com/NVIDIA/apex

  kexinyu

  committed
  ac4ef2d

  Copy full SHA for ac4ef2d

Commits on Apr 30, 2020

• enable wider load/store for multi_tensor_apply kernels (#763)

  Show description for 17ee854
  FDecaYed

  authored
  17ee854

  Copy full SHA for 17ee854

Commits on Apr 28, 2020

• LAMB: global grad clipping & more flexibility in adaptive lr

  kexinyu

  committed
  5b30011

  Copy full SHA for 5b30011

Commits on Sep 6, 2019

• Fix for #456 (#477)

  Show description for 325f5a0
  

  mcarilli

  authored
  325f5a0

  Copy full SHA for 325f5a0

Commits on Aug 16, 2019

• clean up variance options support by all fused optimizers:

  Show description for 18062b6
  FDecaYed

  committed
  18062b6

  Copy full SHA for 18062b6

• add fused lamb, put lamb kernels into one file

  FDecaYed

  committed
  c8f9cce

  Copy full SHA for c8f9cce

Loading