[diffusion] chore: automatically enable dit_layerwise_offload for Wan#16499
[diffusion] chore: automatically enable dit_layerwise_offload for Wan#16499
Conversation
|
/tag-and-rerun-ci |
Summary of ChangesHello @mickqian, 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 streamlines the configuration for "Wan" diffusion models by automatically activating layer-wise CPU offloading when "SGLANG_CACHE_DIT_ENABLED" is not in use. This change aims to improve the out-of-the-box performance for these models by intelligently applying an optimization strategy, while also refining how environment variables are accessed within the system. Highlights
🧠 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 AssistThe 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
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 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
|
There was a problem hiding this comment.
Code Review
This pull request introduces an automatic optimization by enabling dit_layerwise_offload for Wan models, which should improve performance. It also refactors the access to the SGLANG_CACHE_DIT_ENABLED environment variable to use a centralized helper.
My review focuses on improving the maintainability of the new logic. I've suggested a more robust way to identify the model type instead of relying on class name inspection. I've also included a stylistic recommendation to improve the readability of multi-line help strings for command-line arguments.
|
Hi! It is fine to enable this feature by default, but please provide an option to override from CLI - now Wan models are forced to use layerwise DiT offloading even if user specifies |
Motivation
Modifications
Accuracy Tests
Benchmarking and Profiling
Checklist
Review Process
/tag-run-ci-label,/rerun-failed-ci,/tag-and-rerun-ci) or contact authorized users to do so.