use fp32 for e_score_correction_bias in GLM-4.5 #8729
Conversation
There was a problem hiding this comment.
Summary of Changes
Hello @zRzRzRzRzRzRzR, 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 focuses on maintaining numerical consistency within the GLM-4.5 Mixture-of-Experts (MoE) model. The core change ensures that a specific bias parameter, e_score_correction_bias, is always represented in FP32, aligning its precision with what was used during the model's training phase. This is a targeted adjustment to prevent potential precision mismatches during inference.
Highlights
- Model Parameter Data Type Enforcement: I've updated the
GLM4MoEModelclass inpython/sglang/srt/models/glm4_moe.pyto explicitly initialize thee_score_correction_biasparameter withtorch.float32. This ensures that this specific bias term always uses single-precision floating-point, regardless of the model's default data type.
Using Gemini Code Assist
The 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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
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 .gemini/ folder in the base of the repository. Detailed instructions can be found here.
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 or fill out our survey to provide feedback.
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
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
Force e_score_correction_bias to FP32 instead of using the model's own dtype. This is to maintain consistency with what was passed during training.