[VLM Multi-turn] delete fsdp

[gxlvera]

This PR delete fsdp-related content for geo3k vlm multi-turn example, and only keeps Megatron-LM as training backend.

[gemini-code-assist]

Summary of Changes

Hello @gxlvera, 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 geo3k_vlm_multi_turn example by removing all FSDP training backend configurations and related code. The primary goal is to simplify the example's setup and execution by standardizing on Megatron-LM as the sole supported training backend, ensuring consistency and reducing maintenance overhead for this specific VLM multi-turn use case.

Highlights

• FSDP Removal: The FSDP (Fully Sharded Data Parallel) training backend option has been completely removed from the geo3k_vlm_multi_turn example.
• Megatron-LM Standardization: The example now exclusively uses Megatron-LM as its training backend, simplifying the configuration and execution flow.
• Configuration Updates: Both the README.md and run_geo3k_vlm_multi_turn.py files have been updated to reflect these changes, including modifying default environment variable values and removing FSDP-specific code and validation.

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 successfully removes the FSDP training backend from the geo3k VLM multi-turn example, leaving Megatron-LM as the sole backend. The changes are straightforward and mainly involve deleting FSDP-related code and configurations. I've provided a few suggestions to further simplify the code by removing now-redundant logic and to address a minor style issue.

I am having trouble creating individual review comments. Click here to see my feedback.

examples/geo3k_vlm_multi_turn/README.md (33)

Following the suggestion to hardcode the TRAIN_BACKEND in run_geo3k_vlm_multi_turn.py, this environment variable declaration becomes obsolete for this example and can be removed to simplify the setup instructions.

examples/geo3k_vlm_multi_turn/run_geo3k_vlm_multi_turn.py (18-19)

Since FSDP support is being removed and only the 'megatron' backend is supported, you can simplify this by making TRAIN_BACKEND a constant. This eliminates the need to read from an environment variable and perform an assertion, making the code cleaner and more direct.

TRAIN_BACKEND = "megatron"

examples/geo3k_vlm_multi_turn/run_geo3k_vlm_multi_turn.py (145-147)

With the removal of this else block, the if TRAIN_BACKEND == 'megatron': condition at line 135 is always true. Consider refactoring by removing the if statement and un-indenting its body to simplify the code.

examples/geo3k_vlm_multi_turn/run_geo3k_vlm_multi_turn.py (162)

The newline character at the end of the file has been removed. According to PEP 8, files should end with a single newline. Please add it back for consistency.

[zhaochenyang20]

Why change this?