Add PP support for dots_vlm

[gty111]

Add PP support for dots_vlm.

Test with TP8xPP2

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.

[gemini-code-assist]

Summary of Changes

Hello @gty111, 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 pipeline parallelism support for the dots_vlm model. The primary goal is to enhance the model's ability to scale and utilize distributed computing resources more effectively. This is achieved by modifying the model's initialization to recognize pipeline groups and adjusting the forward pass to handle data flow conditionally, depending on whether the current rank is the first in the pipeline. This change is crucial for improving the performance and scalability of large multimodal models.

Highlights

• Pipeline Parallelism Integration: Introduced support for pipeline parallelism (PP) within the dots_vlm model by integrating get_pp_group and PPProxyTensors.
• Conditional Forward Pass: Modified the forward method to execute different logic based on the pipeline parallelism rank, ensuring proper data handling across distributed stages.
• Distributed Group Initialization: Initialized the pp_group within the model's constructor to manage distributed operations.

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. ↩

[gty111]

@ShangmingCai @mickqian

[gemini-code-assist]

Code Review

This pull request adds pipeline parallelism (PP) support for the dots_vlm model. The changes are concentrated in DotsVLMForCausalLM, where the forward method is updated to differentiate between the first pipeline rank and subsequent ranks. The first rank handles multimodal embedding, while other ranks receive hidden states from the previous stage. This is a standard and correct approach for enabling PP in multimodal models. The implementation is clean and follows existing patterns in the codebase. The changes look correct and well-implemented.

[ShangmingCai]

LGTM