refedez_t 0.0.2

Tool for automatizing the deployment of easy federated learning examples

ReFedEz

ReFedEz 🚀 is a Python application and library designed to simplify the implementation and deployment of federated learning architectures. It provides a command-line interface (CLI) for deploying servers and clients directly in their target environments, ensuring consistency and reproducibility, and a Python library that seamlessly integrates into your machine learning code, enabling federated learning to work "like magic" with minimal modifications.

Federated learning is a powerful technique for training machine learning models across distributed data sources while maintaining privacy. ReFedEz serves as the "fast.ai of federated learning" – a beginner-friendly framework that prioritizes simplicity and rapid prototyping. It abstracts the underlying complexities, allowing researchers and developers to focus on their ML innovations rather than infrastructure challenges.

Demo 🎥

Experience ReFedEz in action:

<script id="asciicast-demo" src="https://asciinema.org/a/demo.js" async></script>

Features ✨

• Simplicity: Deploy federated learning setups with ease, and adapt it with minimal code changes.
• Multi-Backend Support: Works with NumPy, PyTorch and TensorFlow.
• Reproducible: Bit by bit reproducible, thanks to nix and uv2nix.
• Multi-Node encrypted by default: Self-signed TLS certificates for the communication between nodes.

Documentation 📚

For detailed guides, API reference, and more, visit the Documentation.

Contributing 🤝

Contributions are welcome! Please see the documentation for guidelines.

License 📄

This project is licensed under the MIT license