Skip to content

Do not store mtp_losses/mtp_acceptance in state params#3130

Open
bvandermoon wants to merge 1 commit intomainfrom
bvandermoon-mtp-params
Open

Do not store mtp_losses/mtp_acceptance in state params#3130
bvandermoon wants to merge 1 commit intomainfrom
bvandermoon-mtp-params

Conversation

@bvandermoon
Copy link
Collaborator

@bvandermoon bvandermoon commented Feb 12, 2026

Description

  • Do not store mtp_losses and mtp_acceptance as part of params. This was causing a checkpoint loading error because they are not stored as part of the checkpoints.
  • Update self attribute names to match pre-NNX checkpoints

FIXES: b/483723849

Tests

Successfully ran the XPK workload in the linked bug with checkpoint loading.

Checklist

Before submitting this PR, please make sure (put X in square brackets):

  • I have performed a self-review of my code. For an optional AI review, add the gemini-review label.
  • I have necessary comments in my code, particularly in hard-to-understand areas.
  • I have run end-to-end tests tests and provided workload links above if applicable.
  • I have made or will make corresponding changes to the doc if needed, including adding new documentation pages to the relevant Table of Contents (toctree directive) as explained in our documentation.

@codecov
Copy link

codecov bot commented Feb 12, 2026

Codecov Report

❌ Patch coverage is 92.85714% with 2 lines in your changes missing coverage. Please review.

Files with missing lines Patch % Lines
src/MaxText/layers/multi_token_prediction.py 92.85% 2 Missing ⚠️

📢 Thoughts on this report? Let us know!

Copy link
Collaborator

@parambole parambole left a comment

Choose a reason for hiding this comment

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

LGTM. Thank you for the quick fix.

@bvandermoon bvandermoon force-pushed the bvandermoon-mtp-params branch from ad257bb to 76cfa15 Compare February 14, 2026 00:03
@bvandermoon bvandermoon force-pushed the bvandermoon-mtp-params branch from 76cfa15 to 18ed9ee Compare February 14, 2026 00:06
Copy link
Collaborator

@parambole parambole left a comment

Choose a reason for hiding this comment

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

Thank you for the fix!

Copy link
Collaborator

@suexu1025 suexu1025 left a comment

Choose a reason for hiding this comment

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

Thanks for the fix!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants