feat: Add support for enabling and configuring msprobe via command-line and config.json

[Vectorwh]

PR type

• Bug Fix
• New Feature
• Document Updates
• More Models or Datasets Support

PR information

🚀 New Feature: msprobe Tool Integration
This Pull Request introduces comprehensive support for the msprobe analysis tool. This integration significantly enhances our debugging capabilities by enabling detailed precision information collection for models.

💡 Key Changes
Command-Line Enablement: The msprobe tool can now be enabled/disabled via the --enable_msprobe launch argument, offering flexible and temporary activation.

Configuration File Support: The tool is configurable through the project's msprobe_config.json file. This allows for persistent and detailed configuration of the msprobe parameters, such as data granularity, output path, and specific collection targets.

Experiment results

Paste your experiment result here(if needed).

[gemini-code-assist]

Summary of Changes

Hello @Vectorwh, 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 a new feature to enhance the debugging capabilities of the training framework by integrating the mindstudio-probe precision debugger. This allows users to enable and configure a specialized tool to analyze model precision during the training steps, providing deeper insights and aiding in the identification of numerical issues. The integration is controlled via new command-line arguments and uses a default JSON configuration file.

Highlights

• New Debugging Tool Integration: This PR introduces the integration of mindstudio-probe, a precision debugging tool, into the framework.
• Configuration Options: New arguments enable_msprobe and msprobe_config have been added to ExtraMegatronArguments to control the debugger's activation and configuration file path.
• Default Configuration File: A default msprobe_config.json file is added, providing initial settings for the mindstudio-probe tool, configured for 'statistics' task.
• Conditional Debugger Execution: The train_step method in base.py is modified to conditionally initialize and manage the PrecisionDebugger based on the enable_msprobe argument.
• Dependency Update: The mindstudio-probe package has been added to the requirements/framework.txt file.

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

The pull request introduces a new dump tool using mindstudio-probe. The changes include adding a configuration file, updating dependencies, and integrating the debugger into the training step. Overall, the feature seems well-intentioned for debugging and statistics. However, there are a few areas that could be improved for robustness and error handling, particularly around the debugger's lifecycle and path management.

[gemini-code-assist]

The msprobe_config is defined with a relative path (./msprobe_config.json). This can lead to issues if the working directory changes or if the script is executed from a different location. Consider resolving this path to an absolute path during initialization or making it configurable to ensure robustness.