use timestep_embedding_cuda from sgl-kernel

[Musisoul]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @Musisoul, 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 introduces a performance enhancement by integrating an optional CUDA-accelerated timestep_embedding function across several core modules. The primary goal is to leverage highly optimized kernel operations for timestep embedding calculations, which can lead to faster execution times in environments equipped with sgl-kernel and CUDA. The implementation includes a robust fallback mechanism, ensuring that the system remains fully functional and stable even if the CUDA-optimized version is not available, by reverting to the standard PyTorch implementation.

Highlights

• CUDA Optimization for Timestep Embedding: Introduced an optional import for a CUDA-optimized timestep_embedding function from the sgl_kernel.elementwise library. This allows for significant performance improvements in timestep embedding calculations when the sgl-kernel is installed and CUDA is available.
• Conditional Fallback Mechanism: Implemented a conditional check to use the timestep_embedding_cuda function if it's successfully imported. If the CUDA version is not available (e.g., sgl-kernel is not installed or CUDA is not configured), the system gracefully falls back to the existing PyTorch-based implementation, ensuring compatibility and stability.
• Broad Application Across Modules: The CUDA-optimized timestep embedding is integrated into key modules including lightx2v/models/networks/hunyuan_video/infer/pre_infer.py, lightx2v/models/schedulers/qwen_image/scheduler.py, and lightx2v/models/schedulers/z_image/scheduler.py, enhancing performance in various parts of the video and image generation pipelines.

🧠 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.

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 introduces an optimization by using a CUDA implementation of timestep_embedding from sgl-kernel when available. This is a good performance enhancement. However, the current implementation introduces code duplication in three separate files. The try-except block for the optional import is identical in all three, and the get_timestep_embedding function is duplicated between two scheduler files. My review highlights these areas of duplication and recommends refactoring to improve code maintainability by following the DRY (Don't Repeat Yourself) principle.