Skip to content

Add autocheckpoint feature to MaxText.#3536

Merged
copybara-service[bot] merged 1 commit intomainfrom
autocheckpoint-v7x
Mar 31, 2026
Merged

Add autocheckpoint feature to MaxText.#3536
copybara-service[bot] merged 1 commit intomainfrom
autocheckpoint-v7x

Conversation

@abhinavclemson
Copy link
Copy Markdown
Collaborator

Description

Add autocheckpoint feature to MaxText.

This change introduces enable_autocheckpoint, which allows Orbax to save a checkpoint when a preemption signal is detected. The checkpoint manager is configured with an AnySavePolicy that includes both preemption-based and fixed-interval saving.

Tests

Tested on tpu v7x with and without autocheckpoint enabled.

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.

PiperOrigin-RevId: 892064416
Copy link
Copy Markdown
Collaborator

@richjames0 richjames0 left a comment

Choose a reason for hiding this comment

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

lgtm

Comment thread src/maxtext/common/checkpointing.py
@copybara-service copybara-service Bot merged commit baec42f into main Mar 31, 2026
104 of 109 checks passed
@copybara-service copybara-service Bot deleted the autocheckpoint-v7x branch March 31, 2026 23:30
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