Skip to content

fix grad norm bug#81

Merged
tastelikefeet merged 2 commits intomodelscope:mainfrom
tastelikefeet:fix/0227-3
Feb 27, 2026
Merged

fix grad norm bug#81
tastelikefeet merged 2 commits intomodelscope:mainfrom
tastelikefeet:fix/0227-3

Conversation

@tastelikefeet
Copy link
Copy Markdown
Collaborator

@tastelikefeet tastelikefeet commented Feb 27, 2026

PR type

  • Bug Fix
  • New Feature
  • Document Updates
  • More Models or Datasets Support

PR information

Write the detail information belongs to this PR.

Experiment results

Paste your experiment result here(if needed).

@gemini-code-assist
Copy link
Copy Markdown
Contributor

gemini-code-assist bot commented Feb 27, 2026

Summary of Changes

Hello, 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 primarily addresses a bug related to gradient normalization by modifying the CrossEntropyLoss calculation. When the loss reduction is set to 'sum', the method now returns both the computed loss and the number of valid labels, enabling accurate scaling for gradient normalization. Additionally, a logging improvement was made to provide better visibility into local class initialization within the twinkle.infra module.

Highlights

  • Gradient Normalization Support: Enhanced the CrossEntropyLoss implementation to return the count of valid labels when the reduction strategy is 'sum', which is crucial for correct gradient normalization in distributed training or specific optimization schemes.
  • Local Initialization Logging: Introduced an informational log message in the twinkle.infra module to alert users when a class is initialized locally without a specified remote_group, suggesting a potential need for remote execution.

🧠 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.

Changelog
  • src/twinkle/infra/init.py
    • Imported get_logger from twinkle.utils and initialized a logger instance.
    • Added an INFO level log message warning when remote_group is None during class initialization, suggesting local execution.
  • src/twinkle/loss/cross_entropy.py
    • Modified the __call__ method to return a tuple (loss, (labels != -100).sum()) when self.reduction is 'sum'.
    • Ensured that only loss is returned for other reduction types.
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 by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

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 pull request 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. 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

  1. 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. ↩

gemini-code-assist[bot]
gemini-code-assist bot reviewed Feb 27, 2026
View reviewed changes
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 aims to fix a gradient normalization bug. The main change is in CrossEntropyLoss, where for reduction='sum', the number of tokens is now returned along with the loss. This is likely necessary for correct gradient normalization during accumulation. However, this introduces an inconsistent return type for the method, which can be a source of bugs. I've suggested a change to make the return type consistent. Another change adds a helpful log message for local initialization of remote-capable classes.

@tastelikefeet tastelikefeet merged commit eb91bbb into modelscope:main Feb 27, 2026
1 of 3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@hjh0119 hjh0119 hjh0119 approved these changes

+1 more reviewer

@gemini-code-assist gemini-code-assist[bot] gemini-code-assist[bot] left review comments

Reviewers whose approvals may not affect merge requirements

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@tastelikefeet @hjh0119