Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
71 lines (56 loc) · 4.57 KB

CONTRIBUTING.md

File metadata and controls

71 lines (56 loc) · 4.57 KB

Contributor Guidelines

Welcome! We're thrilled you're interested in contributing to Koina. These guidelines are designed to help you understand how you can contribute effectively and ensure a smooth collaboration.

Getting Started

  • Familiarize Yourself: Make sure to checkout the main README, publication and website.
  • Check the Issue Tracker: Look for open issues labeled “good first issue” or “help wanted” to find tasks suited for beginners.
  • Ask Questions: If you're unsure about anything, don't hesitate to contact us.

Contribution Types

We welcome contributions in various forms, with a special emphasis on:

  • Documentation Models: Improving documentation to help users better understand and evaluate models.
    • Documentation Questionnaire: We've created a documentation questionnaire based on the DOME documentation guidelines. Following this guide for writing the documentation of a contributed model is encouraged.
    • Clarity and Detail: Ensure documentation is clear, concise, and provides all necessary information for users to understand the model's purpose, limitations, and usage.
  • New Models: Contributing new machine learning (ML) models to enhance accessibility and impact. Check out the this guide to get started!

Other contribution types include:

  • Code: Bug fixes, feature implementations, or optimizations.
  • Design: UI/UX improvements, graphics, or branding.
  • Testing: Reporting bugs, writing test cases, or improving test coverage.
  • Ideas: Suggesting new features or improvements.

Quality Control

We don't evaluate models made available on Koina based on their prediction accuracy/performance. We invite all authors of ML models to contribute their models to Koina to improve their impact and accessibility.

Validation

We heavily encourage you to provide test data for their models.

  • Test Data: Providing test data ensures that predictions are not changing due to unexpected effects of dependencies, infrastructure or optimizations.
  • Testing Schema: We've set up a testing schema to ensure the long-term maintainability of models.

Versioning

Older versions of models should stay available to enable users to reproduce past results.

  • Non-Prediction-Affecting Updates: Updates that don't affect the generated predictions, such as performance improvements in pre- and post-processing scripts, don't require a separate version of the model.
  • Versioning Mechanisms:
    • Primary Method: Use numbered folders in the model directory. This ensures all versions of a model are freely available via Koina. This method is limited to changes that don't require the config file to be adjusted.
    • Secondary Method: For changes requiring adjustments to the config file, create a separate version of the model with an incrementing numbered suffix (<model name>_<version number>).

Contribution Process

  1. Fork the Repository: Create a fork of the project repository.
  2. Create a Branch: Use a descriptive branch name (e.g., feature/add-login or fix/bug-123).
  3. Make Your Changes: Follow the project's coding standards and conventions.
  4. Test Your Changes: Ensure your changes work as expected and don't introduce new issues.
  5. Submit a Pull Request (PR):
    • Provide a clear and concise description of your changes.
    • Reference related issues (e.g., “Closes #123”).
    • Be responsive to feedback during the review process.

Code Style and Standards

  • Follow the project's coding conventions (e.g., indentation, naming, etc.).
  • Include comments where necessary to explain complex logic.
  • Ensure your code is linted and formatted according to the project's guidelines.

Reporting Issues

When reporting bugs or suggesting improvements:

  • Open issues on GitHub
  • Include steps to reproduce the issue, expected behavior, and actual behavior.
  • Provide details about your environment (e.g., OS, browser, version).

Communication

  • Be respectful and inclusive in all interactions.
  • Keep discussions focused and constructive.

Licensing

By contributing, you agree to license your work under the project's open-source license (Apache 2.0).

Thank you!

Thank you for your interest in contributing to Koina! Your efforts help make this project better for everyone.

2025-03-18