[intel] skip xpu fbgemm fp8

[leizhenyuan]

skip xpu fbgemm fp8 as a work around, will change after verify xpu fbgemm fp8 status

[gemini-code-assist]

Summary of Changes

Hello @leizhenyuan, 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 implements a temporary workaround for FP8 quantization support verification within unsloth/models/_utils.py. It modifies the verify_fp8_support_if_applicable function by moving the CUDA-specific device capability checks for fbgemm_fp8 and fp8 quantization into a conditional block that only executes if DEVICE_TYPE is "xpu". This change effectively bypasses these checks for non-XPU devices and marks the XPU-specific logic as a placeholder requiring future verification and adjustment.

Highlights

• FP8 Support Check Scope Adjustment: The existing checks for FP8 quantization support, which previously applied to CUDA devices and utilized torch.cuda.get_device_capability(), are now enclosed within an if DEVICE_TYPE == "xpu": block. This means these specific checks are now only executed if the device type is XPU, and are skipped for other device types like CUDA.
• Temporary XPU FBGEMM FP8 Workaround: This modification serves as a temporary workaround to address fbgemm_fp8 support for XPU devices. A todo comment has been added, indicating that the current logic, which attempts to use torch.cuda.get_device_capability() for XPU, needs further verification and future refinement once the actual XPU FBGEMM FP8 status is confirmed.

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 aims to add a workaround to skip fbgemm fp8 support checks for Intel XPU devices. However, the implementation contains a critical issue where the CUDA-specific capability checks are incorrectly placed inside a condition for XPU devices (DEVICE_TYPE == "xpu"). This will cause a crash on XPU systems and also disables the checks for CUDA devices, which is a regression. I've suggested a fix to change the condition to DEVICE_TYPE == "cuda", which restores the intended logic for CUDA devices and correctly skips the check for XPU.