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.

Sign up for GitHub

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

Jump to bottom

Add llama3.3-70b-instruct model configs, tests and HF golden logits #1282

Merged
merged 1 commit into from
Feb 27, 2025
Merged

Add llama3.3-70b-instruct model configs, tests and HF golden logits #1282

merged 1 commit into from
Feb 27, 2025
+117 −0

Conversation

hengtaoguo
Copy link
Collaborator

@hengtaoguo hengtaoguo commented Feb 18, 2025 •
edited
Loading

Description

Add support for llama3.3-70b-instruct for maxtext. This model is of the same architecture of llama3.1-70b.

  • Upload golden logits generated from HuggingFace.
  • 1_test_llama3.3_70b.sh for downloading HuggingFace checkpoings from GCS and convert to Maxtext format.
  • 2_test_llama3.3_70b.sh for decoding, training and golden logits comparison. Including pre-converted checkpoints.
  • Add llama3.3_70b to llama checkpoint conversion.

FIXES: b/395908473

Tests

2_test: KL divergence below threshold (screenshot).

Checklist

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

  • I have performed a self-review of my code.
  • 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.

@hengtaoguo hengtaoguo marked this pull request as ready for review February 19, 2025 00:38
@hengtaoguo hengtaoguo changed the title Add llama3.3-70b-instruct model configs and HF golden logits Add llama3.3-70b-instruct model configs, tests and HF golden logits Feb 20, 2025
@hengtaoguo hengtaoguo force-pushed the hengtaoguo-llama33 branch 2 times, most recently from f32f186 to 531207c Compare February 26, 2025 18:39
@copybara-service copybara-service bot merged commit 6d861d4 into main Feb 27, 2025
14 checks passed
@copybara-service copybara-service bot deleted the hengtaoguo-llama33 branch February 27, 2025 01:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@khatwanimohit khatwanimohit khatwanimohit approved these changes

@gobbleturk gobbleturk Awaiting requested review from gobbleturk gobbleturk is a code owner

@bvandermoon bvandermoon Awaiting requested review from bvandermoon bvandermoon is a code owner

@vipannalla vipannalla Awaiting requested review from vipannalla vipannalla is a code owner

@RissyRan RissyRan Awaiting requested review from RissyRan RissyRan is a code owner

@richjames0 richjames0 Awaiting requested review from richjames0 richjames0 is a code owner

@rni418 rni418 Awaiting requested review from rni418 rni418 is a code owner

@gagika gagika Awaiting requested review from gagika gagika is a code owner

Assignees
No one assigned
Labels
pull ready
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@hengtaoguo @khatwanimohit