[Doc][TPU] Add models and features supporting matrix. #20230
Conversation
|
👋 Hi! Thank you for contributing to the vLLM project. 💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels. Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging. To run CI, PR reviewers can either: Add 🚀 |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Summary of Changes
Hello @QiliangCui, 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 focuses on enhancing the project's documentation to accurately reflect the current support for Google's Tensor Processing Units (TPUs). It introduces a new dedicated section for hardware-specific model compatibility and expands the existing feature compatibility matrix to include TPU hardware, providing clearer guidance for users on supported models and features.
Highlights
- Feature Compatibility Matrix Expansion: The
features/compatibility_matrix.mdfile has been updated to include a new column for 'TPU' hardware, detailing the support status for various features (e.g., chunked-prefill, LoRA, CUDA graph, pooling models) on TPUs. - New TPU Model Support Document: A new documentation page,
models/hardware_supported_models/tpu.md, has been created. This page provides a comprehensive table listing numerous text-only language models (e.g., Mixtral, Llama, Phi, Gemma, Qwen) and their specific support status (fully, partially, or not supported) when running on TPUs.
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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| 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 issue 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 is currently in preview and 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 to provide feedback.
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
-
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. ↩
There was a problem hiding this comment.
Code Review
This pull request updates the documentation to include TPU support for various models and features. The changes are well-structured and add valuable information. I've provided a couple of suggestions to improve the formatting and correctness of the new markdown tables, which will enhance their readability and maintainability.
mgoin
changed the title
|
Hi @yaochengji , could you take a look? Thank you! |
|
|
||
| # TPU Supported Models | ||
| ## Text-only Language Models | ||
|
|
There was a problem hiding this comment.
Define the symbols? What does 🟨 mean?
QiliangCui
force-pushed
the
docs
branch
2 times, most recently
from
9593f0e to
b3c7a4c
Compare
|
Signed-off-by: Qiliang Cui <[email protected]>
Purpose
Update docs to reflect TPU supporting models and features.
Test Plan
use
mkdocs serveand review.Result
Navigation panel:
Expand:
Feature Support Matrix Look:
