[Doc] Clearer mistral3 and pixtral model support description #20926
Conversation
Signed-off-by: Isotr0py <2037008807@qq.com>
|
👋 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 @Isotr0py, 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 enhances the clarity of the supported_models.md documentation by providing more precise descriptions for Mistral3 and Pixtral model support. It differentiates between models supported via their Hugging Face Transformers format and those supported via their native Mistral format implementations within different generation classes, ensuring users can more easily identify the correct architecture for their specific model variant.
Highlights
- Clarified Mistral3 Support: The
Mistral3ForConditionalGenerationentry in the documentation now explicitly states it supports Mistral3 models in 'HF Transformers' format, providing clearer guidance on its intended use. - Clarified Pixtral Support: The
PixtralForConditionalGenerationentry has been updated to indicate support for both Mistral 3 and Pixtral models when they are in their native 'Mistral format'. This update also includes a more accurate example model ID for Pixtral. - Expanded Llava Support Description: The
LlavaForConditionalGenerationentry now includes a note that it also supports Pixtral models when they are in 'HF Transformers' format, addingmistral-community/pixtral-12bas a relevant example model.
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
The pull request clarifies the support for Mistral3 and Pixtral models, differentiating between the Hugging Face Transformers format and the native Mistral format. The documentation updates in docs/models/supported_models.md are clear and improve understanding. A minor suggestion addresses a potential ambiguity in example model listings.
| | `Phi3VForCausalLM` | Phi-3-Vision, Phi-3.5-Vision | T + I<sup>E+</sup> | `microsoft/Phi-3-vision-128k-instruct`, `microsoft/Phi-3.5-vision-instruct`, etc. | | ✅︎ | ✅︎ | | ||
| | `Phi4MMForCausalLM` | Phi-4-multimodal | T + I<sup>+</sup> / T + A<sup>+</sup> / I<sup>+</sup> + A<sup>+</sup> | `microsoft/Phi-4-multimodal-instruct`, etc. | ✅︎ | ✅︎ | ✅︎ | | ||
| | `PixtralForConditionalGeneration` | Pixtral | T + I<sup>+</sup> | `mistralai/Mistral-Small-3.1-24B-Instruct-2503`, `mistral-community/pixtral-12b`, etc. | | ✅︎ | ✅︎ | | ||
| | `PixtralForConditionalGeneration` | Mistral 3 (Mistral format), Pixtral (Mistral format) | T + I<sup>+</sup> | `mistralai/Mistral-Small-3.1-24B-Instruct-2503`, `mistralai/Pixtral-12B-2409`, etc. | | ✅︎ | ✅︎ | |
There was a problem hiding this comment.
The model mistralai/Mistral-Small-3.1-24B-Instruct-2503 is listed as an example for both Mistral3ForConditionalGeneration (HF format) and PixtralForConditionalGeneration (Mistral format). To avoid ambiguity, suggest removing it from the PixtralForConditionalGeneration entry.
| | `PixtralForConditionalGeneration` | Mistral 3 (Mistral format), Pixtral (Mistral format) | T + I<sup>+</sup> | `mistralai/Mistral-Small-3.1-24B-Instruct-2503`, `mistralai/Pixtral-12B-2409`, etc. | | ✅︎ | ✅︎ | | |
| | `PixtralForConditionalGeneration` | Mistral 3 (Mistral format), Pixtral (Mistral format) | T + I<sup>+</sup> | `mistralai/Pixtral-12B-2409`, etc. | | ✅︎ | ✅︎ | |
| | `MiniCPMV` | MiniCPM-V | T + I<sup>E+</sup> + V<sup>E+</sup> | `openbmb/MiniCPM-V-2` (see note), `openbmb/MiniCPM-Llama3-V-2_5`, `openbmb/MiniCPM-V-2_6`, etc. | ✅︎ | | ✅︎ | | ||
| | `MiniMaxVL01ForConditionalGeneration` | MiniMax-VL | T + I<sup>E+</sup> | `MiniMaxAI/MiniMax-VL-01`, etc. | | ✅︎ | ✅︎ | | ||
| | `Mistral3ForConditionalGeneration` | Mistral3 | T + I<sup>+</sup> | `mistralai/Mistral-Small-3.1-24B-Instruct-2503`, etc. | ✅︎ | ✅︎ | ✅︎ | | ||
| | `Mistral3ForConditionalGeneration` | Mistral3 (HF Transformers) | T + I<sup>+</sup> | `mistralai/Mistral-Small-3.1-24B-Instruct-2503`, etc. | ✅︎ | ✅︎ | ✅︎ | |
There was a problem hiding this comment.
| | `Mistral3ForConditionalGeneration` | Mistral3 (HF Transformers) | T + I<sup>+</sup> | `mistralai/Mistral-Small-3.1-24B-Instruct-2503`, etc. | ✅︎ | ✅︎ | ✅︎ | | |
| | `Mistral3ForConditionalGeneration` | Mistral 3 (HF Transformers) | T + I<sup>+</sup> | `mistralai/Mistral-Small-3.1-24B-Instruct-2503`, etc. | ✅︎ | ✅︎ | ✅︎ | |
github-actions
bot
added
the
ready
…oject#20926) Signed-off-by: Isotr0py <2037008807@qq.com> Signed-off-by: x22x22 <wadeking@qq.com>
…oject#20926) Signed-off-by: Isotr0py <2037008807@qq.com>
…oject#20926) Signed-off-by: Isotr0py <2037008807@qq.com>
…oject#20926) Signed-off-by: Isotr0py <2037008807@qq.com> Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>
…oject#20926) Signed-off-by: Isotr0py <2037008807@qq.com> Signed-off-by: Paul Pak <paulpak58@gmail.com>
…oject#20926) Signed-off-by: Isotr0py <2037008807@qq.com> Signed-off-by: Diego-Castan <diego.castan@ibm.com>
…oject#20926) Signed-off-by: Isotr0py <2037008807@qq.com>
4 participants
Essential Elements of an Effective PR Description Checklist
supported_models.mdandexamplesfor a new model.Purpose
docs/models/supported_models.mdwith HF/Mistral format description about mistral3/pixtral.Test Plan
Test Result
(Optional) Documentation Update