Add sft tools example and tests#210
Conversation
![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 @garyzhang99, 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 significantly enhances the training framework by introducing capabilities for Supervised Fine-Tuning (SFT) of models that can leverage external tools. It includes the addition of a new dataset specifically designed with tool-use examples, alongside necessary modifications to the data reading and processing pipeline to correctly handle tool definitions and calls. A new test case has also been added to ensure the robustness and correctness of this new functionality, paving the way for training more capable and interactive language models.
Highlights
- Tool-Augmented SFT Support: This PR introduces support for Supervised Fine-Tuning (SFT) of models that interact with external tools. This is achieved by adding a new dataset format that includes tool definitions and tool call examples within conversation messages.
- New Tool-Use Example Dataset: A new example dataset,
sft_with_tools.json, has been added. This dataset contains diverse scenarios where a language model uses tools like weather lookups, calculators, web search, email sending, and currency conversion, providing rich training data for tool-use capabilities. - Data Pipeline Enhancements for Tools: Core data processing components, including
FileReader,FormatConfig, andExperiencedata structures, have been updated to correctly parse, store, and utilize tool information during the SFT process. This ensures that tool definitions and calls are properly integrated into the training pipeline. - Comprehensive Testing for Tool SFT: A dedicated unit test,
TestTrainerToolsSFT, has been implemented to validate the end-to-end functionality of the SFT trainer with tool-augmented data, ensuring that the new features work as expected and maintain system stability.
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 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
-
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. ↩
|
/unittest-module-trainer |
There was a problem hiding this comment.
Code Review
This pull request adds support for tool usage in Supervised Fine-Tuning (SFT). The changes include adding a new test dataset with tool calls, updating the data reader to handle the tools field, and adding a new integration test for SFT with tools. The changes are logical and well-structured. I have a few suggestions to improve code clarity and maintainability.
Summary
Failed Tests
Tests
Github Test Reporter by CTRF 💚 |
|
/unittest-module-trainer |
Summary
Tests
Github Test Reporter by CTRF 💚 |
|
/unittest-module-trainer |
Summary
Tests
Github Test Reporter by CTRF 💚 |
2 participants
Description
As the title says.
Checklist
Please check the following items before code is ready to be reviewed.