Skip to content

Conversation

xibinliu
Copy link
Collaborator

@xibinliu xibinliu commented Oct 7, 2025

Add the following new training recipes for v5p:

  • DeepSeek-671B-MaxText
  • Llama3.1-405B-MaxText

Also updated the Llama4-Scout-17B-16E Maxtext recipe to use JAX 0.7.0 and newer code commit.

The README files under training/trillium has been moved to the upper level folder and the reference links are updated.

TODO: Update the commit when python versioning PR is ready.

Copy link
Contributor

@RissyRan RissyRan left a comment

Choose a reason for hiding this comment

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

LGTM! @bvandermoon may have up-to-date info about versioning part.

@xibinliu xibinliu force-pushed the xibin/v5p_recipes branch 2 times, most recently from 16b3946 to 378a9d0 Compare October 8, 2025 18:00
@xibinliu xibinliu merged commit 7e88b6d into main Oct 8, 2025
1 check passed
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