The implementation of federated learning PyTorch respectively.

1.Python 3.9.13

2.pytorch 1.3.1

both of them run on GPU

You are supposed to prepare the data set by yourself. MNIST can be downloaded on http://yann.lecun.com/exdb/mnist/. These data sets should be put into /data/MNIST when the download is finished.

Run the code

python server.py -nc 100 -cf 1 -E 5 -B 128  -mn mnist_2nn  -ncomm 50 -iid 0 -lr 0.001 -vf 1 -g 1 -dp 0.5 -op SGD -poipro 0.5 -revprob 0.5 -threshold 0.8 -repoi trainBehavior

which means there are 100 clients. The data set are allocated in Non-IID way. The epoch and batch size are set to 5 and 128. The learning rate is 0.01, we validate the codes every 20 rounds during the training, training stops after 50 rounds. There are three models to do experiments: mnist_2nn mnist_cnn and cifar_cnn, and we choose mnist_cnn in this command. Notice the data set path when run the code of pytorch-version(you can take the source code out of the 'use_pytorch' folder).