Enhance support for VL models #501
Open
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
…om the sample client. 2. Updated the `SFTFormatter`, `vLLMRolloutModel`, `SimpleMMWorkflow`, `mm_utils.py`, `verl_trainer.py` and `trainer/verl/utils.py` to enhance multimodal processing capabilities, enabling compatibility with Qwen-series vision-language models.
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 significantly enhances the framework's support for Vision-Language (VL) models, particularly the Qwen-series. The changes streamline how multimodal inputs are processed and handled throughout the system, from data formatting and model interaction to training. By integrating multimodal capabilities directly into core Highlights
Changelog
Activity
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 significantly enhances support for Vision-Language models by refactoring the multi-modal data processing pipeline. The changes centralize logic, unify text and multi-modal generation paths, and simplify higher-level components like workflows, which is a great architectural improvement. However, I've identified a critical bug due to a missing import, along with some high-severity concerns regarding potential data loss of attention_mask and inadequate error handling that could lead to an IndexError. Please review the detailed comments for suggestions on how to address these issues.
2. Fix `GPUMemoryValidator` for VL model
2. Remove `pad_token_id` in config. 3. Add `trust_remote_code` to config. 4. Add `get_model_class` for fsdp worker and fsdp checkpoint manager.
chenyushuo
changed the title
Collaborator
Author
|
/unittest-module-trainer |
Summary
Failed Tests
Skipped
Tests
Github Test Reporter by CTRF 💚 |
Collaborator
Author
|
/unittest-module-trainer |
Summary
Failed Tests
Skipped
Tests
Github Test Reporter by CTRF 💚 |
Collaborator
Author
|
/unittest-module-trainer |
Summary
Skipped
Tests
Github Test Reporter by CTRF 💚 |
pan-x-c
reviewed
Feb 11, 2026
garyzhang99
reviewed
Feb 11, 2026
pan-x-c
reviewed
Feb 11, 2026
Collaborator
Author
|
/unittest-all |
Collaborator
Author
|
/unittest-module-algorithm |
Collaborator
Author
|
/unittest-module-buffer |
Collaborator
Author
|
/unittest-module-cli |
Collaborator
Author
|
/unittest-module-common |
Collaborator
Author
|
/unittest-module-explorer |
Collaborator
Author
|
/unittest-module-manager |
Collaborator
Author
|
/unittest-module-service |
Collaborator
Author
|
/unittest-module-utils |
Collaborator
Author
|
/unittest-module-trainer |
Summary
Tests
Github Test Reporter by CTRF 💚 |
Summary
Tests
Github Test Reporter by CTRF 💚 |
Summary
Tests
Github Test Reporter by CTRF 💚 |
Summary
Skipped
Tests
Github Test Reporter by CTRF 💚 |
Summary
Tests
Github Test Reporter by CTRF 💚 |
Summary
Tests
Github Test Reporter by CTRF 💚 |
Summary
Tests
Github Test Reporter by CTRF 💚 |
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
3 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
Modified the tokenizer in
TinkerModelto directly retrieve it from the sample client.Updated the
SFTFormatter,vLLMRolloutModel,SimpleMMWorkflow,mm_utils.py,verl_trainer.pyandtrainer/verl/utils.pyto enhance multimodal processing capabilities, enabling compatibility with Qwen-series vision-language models andmoonshotai/Kimi-VL-A3B-Thinking.Remove
pad_token_idin config.Add
trust_remote_codeto config.Add
get_model_classfor fsdp worker and fsdp checkpoint manager.Bug fix in
chord_policy_loss.py.Checklist
Please check the following items before code is ready to be reviewed.