Federated Learning Docker on RPI Cluster

Reference

@article{beutel2020flower,
  title={Flower: A Friendly Federated Learning Research Framework},
  author={Beutel, Daniel J and Topal, Taner and Mathur, Akhil and Qiu, Xinchi and Parcollet, Titouan and Lane, Nicholas D},
  journal={arXiv preprint arXiv:2007.14390},
  year={2020}
}

Configuration

Model

The deault model is MobileNet

• Modify the global model at line 225 of src/server/server.py
• Modify the local model at line 17 of src/client.py

Dataset

The default datasets is CIFAR10

• Modify the dataset at line 19 of src/client.py The default dataset distribution is randomly picking up
• Modify dataset distribution at sampling_data() method of src/client.py

Federated Learning algorithm

The default federated learning algorithm is FedAvg

• Modify the default federated learning algorithm at line 63 of src/server/server.py

SERVER

BUILD

sudo docker build -t server -f Dockerfile.server .

RUN

sudo docker run -d -p 19191:19191 -v global_models:/global_models kundjanasith/pik8ke_fl:server

CLIENT

BUILD

sudo docker build -t client -f Dockerfile.client .

RUN

sudo docker run -d -e server=<server_ip> -e client=<client_ip> -v local_models:/local_models kundjanasith/pik8ke_fl:client