Add local LCB evaluation

[nv-alicheng]

What does this PR do?

Adds LCB evaluation functionality via LCBServe. Followup PR will move LCBServe to a containerized microservice.

Type of change

• Bug fix
• New feature
• Documentation update
• Refactor/cleanup

Related issues

Testing

• Tests added/updated
• All tests pass locally
• Manual testing completed

Checklist

• Code follows project style
• Pre-commit hooks pass
• Documentation updated (if needed)

[github-actions]

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

[gemini-code-assist]

Summary of Changes

Hello @nv-alicheng, 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 evaluation capabilities by incorporating LiveCodeBench (LCB) into the SGLang framework. It establishes a robust mechanism for evaluating code generation models by introducing a dedicated service for secure code execution, a specialized extractor for Python code, and seamlessly integrating LCB into the comprehensive accuracy benchmark suite. This lays the groundwork for more rigorous and secure code-based model assessments.

Highlights

• LiveCodeBench (LCB) Integration: Introduced functionality to evaluate models against the LiveCodeBench dataset, a benchmark for code generation tasks.
• LCBServe Component: Added a new LCBServe module designed to run LCB evaluation in a controlled, isolated environment, with a future plan for containerization.
• Python Code Extraction: Implemented a PythonCodeExtractor to accurately extract Python code from markdown blocks in model outputs, crucial for LCB evaluation.
• Benchmark Suite Expansion: Integrated LCB evaluation into the existing SGLang Accuracy Suite Benchmark, alongside GPQA and AIME25, and updated related scripts and report naming conventions.
• LCB Setup Script: Provided a setup_lcb.sh script to streamline the cloning and dependency installation process for the LiveCodeBench repository.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

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 adds support for LiveCodeBench (LCB) evaluation. The changes include a new evaluation script for LCB, a setup script, the core evaluation logic in lcb_serve.py, a new PythonCodeExtractor, and a LiveCodeBenchScorer. The main benchmark runner is also updated to include LCB. My review focuses on improving robustness, configurability, and correctness. I've identified several hardcoded paths that should be made configurable, a bug where a timeout parameter is ignored, and some opportunities for code improvement and making the implementation less brittle.

[arekay-nv]

Great work - thanks!