Skip to content

[fix]: fix seq_len pad len, and adapt to new mtp_loss api (for megatron dev brance)#6206

Open
zpltys wants to merge 1 commit intoverl-project:mainfrom
meituan-search:zs-fix
Open

[fix]: fix seq_len pad len, and adapt to new mtp_loss api (for megatron dev brance)#6206
zpltys wants to merge 1 commit intoverl-project:mainfrom
meituan-search:zs-fix

Conversation

@zpltys
Copy link
Copy Markdown
Contributor

@zpltys zpltys commented Apr 29, 2026

  1. when both enable CP and TP, seq_len should pad to (2 * cp) * tp
  2. adapt to new mtp_loss api (for megatron dev brance)

What does this PR do?

Bugfix:
1. when both enable CP and TP, seq_len should pad to (2 * cp) * tp
2. adapt to new mtp_loss api (for megatron dev brance)

Add concise overview of what this PR aims to achieve or accomplish. Reference related GitHub issues and PRs that help with the review.

Checklist Before Starting

  • Search for similar PRs. Paste at least one query link here: ...
  • Format the PR title as [{modules}] {type}: {description} (This will be checked by the CI)
    • {modules} include fsdp, megatron, veomni, sglang, vllm, rollout, trainer, ci, training_utils, recipe, hardware, deployment, ray, worker, single_controller, misc, perf, model, algo, env, tool, ckpt, doc, data, cfg, reward, fully_async, one_step_off
    • If this PR involves multiple modules, separate them with , like [megatron, fsdp, doc]
    • {type} is in feat, fix, refactor, chore, test
    • If this PR breaks any API (CLI arguments, config, function signature, etc.), add [BREAKING] to the beginning of the title.
    • Example: [BREAKING][fsdp, megatron] feat: dynamic batching

Test

For changes that can not be tested by CI (e.g., algorithm implementation, new model support), validate by experiment(s) and show results like training curve plots, evaluation results, etc.

API and Usage Example

Demonstrate how the API changes if any, and provide usage example(s) if possible.

# Add code snippet or script demonstrating how to use this

Design & Code Changes

Demonstrate the high-level design if this PR is complex, and list the specific changes.

Checklist Before Submitting

Important

Please check all the following items before requesting a review, otherwise the reviewer might deprioritize this PR for review.

1. when both enable CP and TP, seq_len should pad to (2 * cp) * tp
2. adapt to new mtp_loss api (for megatron dev brance)
@zpltys zpltys changed the title Bugfix: fix seq_len pad len, and adapt to new mtp_loss api (for megatron dev brance) [fix]: fix seq_len pad len, and adapt to new mtp_loss api (for megatron dev brance) Apr 29, 2026
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

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

Code Review

This pull request updates the Multi-Token Prediction (MTP) loss calculation to support the new Megatron API while maintaining a legacy fallback. It also adjusts the sequence length alignment logic in utility functions to ensure compatibility with zigzag Context Parallel (CP) and Tensor Parallel (TP) configurations by calculating align_size as tp_size * cp_size * 2. I have no feedback to provide.

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.

1 participant