[FIX] Move labels to logits device in cross-entropy loss for multi-GPU support (#4041)

[devchilll]

When using device_map='balanced' with multiple GPUs, the labels tensor may reside on a different device than the logits/losses tensors. This causes a RuntimeError at the masked_fill_ call in the chunked cross-entropy forward path.

Fix: explicitly move labels to the same device as logits at the start of Fast_CrossEntropyLoss.forward(). This is a no-op on single-GPU setups.

Fixes #4041

[gemini-code-assist]

Summary of Changes

Hello @devchilll, 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 resolves a RuntimeError in multi-GPU configurations by ensuring device consistency for label tensors during cross-entropy loss calculations. The change explicitly moves labels to the same device as logits, preventing device mismatches and stabilizing the loss computation without impacting single-GPU setups.

Highlights

• Multi-GPU Compatibility Fix: Implemented a fix to ensure that label tensors are explicitly moved to the same device as logits within the Fast_CrossEntropyLoss.forward() method. This resolves a RuntimeError encountered in multi-GPU environments when device_map='balanced' is used, where labels might be on a different device than logits.

🧠 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

• unsloth/kernels/cross_entropy_loss.py
  • Added a line to explicitly move the labels tensor to the same device as the logits tensor at the beginning of the Fast_CrossEntropyLoss.forward() method.

Activity

• No human activity has been recorded on this pull request yet.

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]

Code Review

This pull request addresses a device mismatch issue that can occur in multi-GPU environments within the Fast_CrossEntropyLoss function. The change correctly moves the labels tensor to the same device as the logits tensor at the beginning of the forward pass. This prevents potential runtime errors during the loss calculation when tensors are on different devices. The fix is concise, idiomatic, and effectively resolves the described problem without any apparent side effects. The change is well-implemented.

[danielhanchen]

Thanks!