Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix device error while calculating loss in Trainer #32480

Closed
wants to merge 1 commit into from

Conversation

echo-yi
Copy link

@echo-yi echo-yi commented Aug 7, 2024

What does this PR do?

Fixes ValueError in Trainer when tr_loss and tr_loss_step are on different devices. This error happens when I use device_map='auto' and ORPOTrainer together.

Before submitting

  • This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
  • Did you read the contributor guideline,
    Pull Request section?
  • Was this discussed/approved via a Github issue or the forum? Please add a link
    to it if that's the case.
  • Did you make sure to update the documentation with your changes? Here are the
    documentation guidelines, and
    here are tips on formatting docstrings.
  • Did you write any new necessary tests?

Who can review?

@amyeroberts

@amyeroberts
Copy link
Collaborator

Hi @echo-yi, thanks for opening this PR!

Based on the warning message in the diff - it seems this is flagging device placement which should be handled elsewhere. cc @muellerzr who might know more

@SunMarc
Copy link
Member

SunMarc commented Aug 12, 2024

Hey @echo-yi, I've opened up a PR to fix the issue you are facing. The fix should live in TRL instead. LMK if this solves your issue !

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.

3 participants