Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
137 lines (97 loc) · 2.66 KB

CONTRIBUTING.md

File metadata and controls

137 lines (97 loc) · 2.66 KB
Raw

Contributing to LLM Fine-Tuning Lab

Thank you for your interest in contributing to the LLM Fine-Tuning Lab! This document provides guidelines and instructions for contributing.

Getting Started

  1. Fork the repository

  2. Clone your fork

    git clone https://github.com/YOUR_USERNAME/LLM-Finetuning-Lab.git
cd LLM-Finetuning-Lab

  3. Set up development environment

    make dev-install

  4. Create a feature branch

    git checkout -b feature/your-feature-name

Development Guidelines

Code Style

We follow PEP 8 style guidelines with some modifications:

  • Maximum line length: 120 characters
  • Use type hints for function signatures
  • Write docstrings for all public functions and classes

Format your code before committing:

make format

Testing

All new features must include tests:

# Run tests
make test

# Run specific test file
pytest tests/test_training.py

Commit Messages

Follow conventional commit format:

<type>(<scope>): <subject>

<body>

<footer>

Types:

  • feat: New feature
  • fix: Bug fix
  • docs: Documentation changes
  • style: Code style changes
  • refactor: Code refactoring
  • test: Adding tests
  • chore: Maintenance tasks

Example:

feat(training): add LoRA fine-tuning support

Implemented LoRA-based fine-tuning for efficient parameter
adaptation. Includes configuration options and example scripts.

Closes #123

Pull Request Process

  1. Update documentation if you're adding new features
  2. Add tests for new functionality
  3. Run linters and ensure all tests pass 
    make lint
make test
  4. Update CHANGELOG.md with your changes
  5. Submit PR with a clear description

PR Checklist

  • Code follows style guidelines
  • Tests added and passing
  • Documentation updated
  • CHANGELOG.md updated
  • Commit messages follow convention

Areas for Contribution

High Priority

  • Implement RLHF training pipeline
  • Add support for multimodal models
  • Optimize distributed training
  • Improve documentation

Medium Priority

  • Add more evaluation metrics
  • Create tutorial notebooks
  • Implement model distillation
  • Add CI/CD improvements

Good First Issues

Look for issues labeled good-first-issue in the issue tracker.

Code Review Process

  1. At least one maintainer review required
  2. All CI checks must pass
  3. Documentation must be updated
  4. Changes must be tested

Questions?

Feel free to:

Thank you for contributing to SynthoraAI!