Skip to content

[GRPO] disabling top_k sampling default #3494

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

Merged
merged 3 commits into from
May 26, 2025
Merged

[GRPO] disabling top_k sampling default #3494

merged 3 commits into from
May 26, 2025

Conversation

kashif
Copy link
Collaborator

@kashif kashif commented May 26, 2025

What does this PR do?

Disabling top_k in GRPO sampling as a default and setting the sampling temperature default to 1.0

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? Please add a link
    to it if that's the case.
  • Did you make sure to update the documentation with your changes?
  • Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@kashif kashif requested a review from lewtun May 26, 2025 08:51
Copy link
Member

@lewtun lewtun left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM with a question about top_k=None and vllm that we should confirm before merging

Copy link
Member

@lewtun lewtun left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM with two nits!

kashif and others added 2 commits May 26, 2025 11:03
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>
@HuggingFaceDocBuilderDev

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

@kashif kashif merged commit 5c08897 into main May 26, 2025
3 of 10 checks passed
@kashif kashif deleted the grpo-defaults branch May 26, 2025 09:32
@qgallouedec
Copy link
Member

For reference, this has been elegantly reported here: https://x.com/kalomaze/status/1926751357983154606

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.

4 participants