Update veRL to 0.7.0 #471
Open
Update veRL to 0.7.0 #471
+1,617
−616
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
Summary of ChangesHello @chenyushuo, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request integrates the latest Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request updates the veRL dependency to version 0.7.0. The changes are extensive, reflecting a major version upgrade. Key modifications include updates to configuration classes in trinity/common/verl_config.py to align with new veRL features, and significant refactoring in worker implementations (fsdp_workers.py, megatron_workers.py) to support new functionalities like advanced profiling, router replay for MoE models, and improved LoRA adapter handling. Checkpointing logic has also been enhanced for better reliability and to support asynchronous operations. Overall, the changes are well-integrated and necessary for the dependency upgrade. I have one minor suggestion to ensure consistency in metric logging.
pan-x-c
reviewed
Jan 15, 2026
Collaborator
Author
|
/unittest-module-trainer |
Summary
Failed Tests
Skipped
Tests
Github Test Reporter by CTRF 💚 |
Collaborator
Author
|
/unittest-all |
Summary
Failed Tests
Skipped
Tests
Github Test Reporter by CTRF 💚 |
Collaborator
Author
|
/unittest-module-explorer |
Summary
Skipped
Tests
Github Test Reporter by CTRF 💚 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
As the title says.
Experimental Results
GSM8k
In the figure below,
test_trainerdenotes training conducted using Fully Sharded Data Parallel (FSDP), whiletest-megatron_trainerrefers to training performed with Megatron.The training curve obtained with Megatron exhibits a constant scaling discrepancy relative to the corresponding FSDP curve, due to the omission of scaling the output
actor/kl_lossby the number of micro-batches (n_micro_batch). This issue has been addressed in the latest code version.All other curves remain consistent with those from previous experiments.
Countdown
In the figure below,
test_trainercorresponds to FSDP-based training, andtest-megatron_trainercorresponds to Megatron-based training.All curves align closely with those observed in prior experiments.
Guru
FSDP
In the figure below,
guru-benchrepresents the experimental results from before this pull request (PR), whereasguru_math-benchreflects the results obtained in this PR.Due to the introduction of scaling for the output
actor/kl_lossin this PR, a constant multiplicative difference is observed compared to historical curves. All other metrics remain consistent with prior experimental results.Megatron
In the figure below,
fixed-moe-is-megatrondenotes the experimental results from before this PR, andfixed-moe-is-megatron-verl-0.7.0corresponds to the results from this PR.As with the FSDP case, the scaling of the output
actor/kl_lossintroduced in this PR leads to a constant multiplicative offset relative to previous curves. Aside from this adjustment, all other curves are consistent with earlier experimental outcomes.Bug Fix for Entropy Calculation in Megatron
The image below illustrates GPU memory utilization: the yellow box corresponds to our current implementation, while the red box shows the memory usage after applying the veRL fix.
The veRL fix incurs higher GPU memory consumption than our implementation—attributable to the use of
.clone()versus a monkey patch applied directly to the Megatron function.Consequently, we have temporarily commented out the veRL fix pending further optimization.
Checklist
Please check the following items before code is ready to be reviewed.