Skip to content

Releases: aws/sagemaker-pytorch-training-toolkit

v3.0.0

25 Aug 20:14
Compare
Choose a tag to compare

Breaking Changes

  • Update sagemaker-training version >=5.0.0

v2.9.0

04 Jun 18:45
Compare
Choose a tag to compare

Features

  • Add Code Owners file

Bug Fixes and Other Changes

  • upper bounding dependency on sagemaker-training-toolkit

v2.8.1

23 Oct 15:20
Compare
Choose a tag to compare

Bug Fixes and Other Changes

  • bypass DNS check for studio local exec

v2.8.0

28 Apr 04:21
Compare
Choose a tag to compare

Features

  • add PyTorch version environment variable, to facilitate SMTT

Bug Fixes and Other Changes

  • pin coverage version to fix pipeline issue

v2.7.0

20 Oct 19:23
Compare
Choose a tag to compare

Features

  • Add torch_distributed support for Trainium

v2.6.2.post0

21 Sep 15:19
Compare
Choose a tag to compare

Documentation Changes

  • update README and contributing guidelines

v2.6.2

26 Aug 17:50
Compare
Choose a tag to compare

Bug Fixes and Other Changes

  • provide option to use native process launcher

v2.6.1

15 Aug 23:30
Compare
Choose a tag to compare
prepare release v2.6.1

v2.6.0

03 Aug 02:07
Compare
Choose a tag to compare

Features

  • add support for native PyTorch DDP distribution

v2.5.1

21 Jul 21:18
Compare
Choose a tag to compare

Bug Fixes and Other Changes

  • deriver master node from training environment