decentralearn 2.1.1

A decentralized federated learning framework with privacy guarantees, created by Jihad GARTI

DecentraLearn

A decentralized federated learning framework with strong privacy guarantees, built on blockchain technology.

🚀 Features

• Federated Learning

  • Decentralized model training
  • Secure model aggregation
  • Incentive mechanisms
  • Smart contract-based coordination

• Privacy Mechanisms

  • Differential Privacy
  • Homomorphic Encryption
  • Zero-Knowledge Proofs
  • Secure Aggregation

• Blockchain Integration

  • Smart contract-based verification
  • Transparent model tracking
  • Decentralized coordination
  • Incentive distribution

• Advanced Security

  • End-to-end encryption
  • Model integrity verification
  • Access control
  • Audit logging

📦 Installation

Install via pip (Recommended)

You can install the latest release (v2.1.0) directly from PyPI:

pip install decentralearn

To upgrade to the latest version:

pip install --upgrade decentralearn

Note: If you encounter dependency issues (especially with eth-tester), try installing the beta version first:

pip install eth-tester==0.13.0b1
pip install decentralearn

Or install all dependencies manually:

pip install torch>=2.0.0 numpy>=1.21.0 scipy>=1.7.0 opacus>=1.1.0 phe>=1.5.0 web3>=6.0.0 eth-tester==0.13.0b1 eth-utils>=2.1.0
pip install decentralearn

---

Prerequisites

• Python 3.8 or higher
• Ethereum node (e.g., Ganache for development)
• PyTorch 1.7 or higher

Installation Steps (from source)

1. Clone the repository:

   git clone https://github.com/ackerman23/decentralearn
   cd decentralearn
   

2. Create and activate a virtual environment:

   python -m venv .venv
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate
   

3. Install dependencies:

   pip install -e .
   

🏁 Quick Start

1. Initialize blockchain connection:

   from decentralearn.blockchain.client import BlockchainClient
   from decentralearn.config.blockchain_config import BlockchainConfig
   
   config = BlockchainConfig(
       rpc_url="http://localhost:8545",
       chain_id=1337
   )
   client = BlockchainClient(config)
   

2. Create and train a model:

   from decentralearn.models.base import BaseModel
   from decentralearn.privacy import DifferentialPrivacy
   
   model = BaseModel()
   dp = DifferentialPrivacy(epsilon=0.1)
   # Train model with privacy
   

3. Upload to blockchain:

   client.upload_model(model)
   

📚 Documentation

Comprehensive documentation is available in the docs/ directory:

• Architecture Overview
• API Reference
• Tutorials
• Examples
• Development Guide

🧪 Testing

Run the test suite:

pytest tests/

For detailed test output:

pytest -v tests/

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

1. Fork the repository
2. Create your feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

Creator: Jihad GARTI ([e-mail])

📧 Support

For questions and support:

1. Check the FAQ
2. Open an issue on GitHub
3. Contact the maintainers

🙏 Acknowledgments

• PyTorch team for the deep learning framework
• Ethereum community for blockchain infrastructure
• Privacy research community for privacy-preserving techniques

📄 Citation

If you use DecentraLearn in your research, please cite:

@software{decentralearn2024,
  author = {Jihad GARTI},
  title = {DecentraLearn: A Decentralized Federated Learning Framework},
  year = {2024},
  publisher = {Jihad GARTI},
  url = {https://github.com/ackerman23/decentralearn}
}