Skip to content

Store training state and resume training from stored state. Fix bug i…#103

Merged
mseeger merged 1 commit into
mainfrom
restart_checkpoint3
May 6, 2026
Merged

Store training state and resume training from stored state. Fix bug i…#103
mseeger merged 1 commit into
mainfrom
restart_checkpoint3

Conversation

@mseeger
Copy link
Copy Markdown
Contributor

@mseeger mseeger commented May 6, 2026

…n SFTDataset.getitem (#103)

  • Provides storing training state and resume training (Support restart from checkpoint #101)
  • Fixed bad bug: SFTDataset.__getitem__ was making random choices with every call. Choices are now made once and stored in the metadata

Closes #101.


By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

@mseeger mseeger merged commit 3b996ee into main May 6, 2026
7 checks passed
@mseeger mseeger deleted the restart_checkpoint3 branch May 6, 2026 13:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Support restart from checkpoint

1 participant