FedEasy: Federated Learning With Ease

FedEasy Logo

FedEasy is an intuitive, powerful, yet simple-to-use federated learning framework built on top of the Flower framework and PyTorch. Our goal is to streamline the process of setting up and running federated learning experiments, making advanced machine learning techniques accessible to researchers and developers alike.

📢 Publication

FedEasy has been peer-reviewed and published in the SoftwareX journal (Elsevier).
Read the full article for more info.

Key Features

  • Flower and PyTorch Integration: Leverage the strengths of both the Flower framework and PyTorch to build a robust federated learning environment.

  • Configurability: Customize your federated learning setup using a single YAML file, allowing you to configure data distributions, models, optimizers, training strategies, and more.

  • Extensive Documentation: Explore detailed explanations of supported datasets, models, and strategies, making it easier to select the right components for your project.

  • Easy Setup: Follow a straightforward setup process using Git, Python, pip, and Miniconda to get your federated learning environment up and running quickly.

  • Flexible Deployment: Run simulations on a single machine or scale effortlessly to multiple clients. You can also deploy in a multi-node setup to handle high computational demands.

  • Performance Metrics: Save and analyze training results, including accuracy, loss, and time per round, to monitor the progress of your federated learning models.

  • Community Contribution: Contribute to the repository by enhancing existing features, adding new models or strategies, or improving documentation. Contributions are welcome and help make federated learning more accessible to everyone.

Note: While FedEasy is primarily tested in Linux environments, Windows users can also use it with some adjustments.

Citation

If you use FedEasy in your research, please cite our work:

@article{KUNDROO2025102276,
title = {FedEasy : Federated learning with ease},
journal = {SoftwareX},
volume = {31},
pages = {102276},
year = {2025},
issn = {2352-7110},
doi = {https://doi.org/10.1016/j.softx.2025.102276},
url = {https://www.sciencedirect.com/science/article/pii/S2352711025002432},
author = {Majid Kundroo and Ghani Haider and Nguyen Khoa and Abdul Wahab Mamond and Taehong Kim},
keywords = {Federated learning, Framework, Federated datasets}
}

Contact

For questions, suggestions, or collaborations, please reach out to:

FedEasy is released under this License.