Skip to content

Use federated learning to train variational auto-encoders on disjoint distributions

Notifications You must be signed in to change notification settings

stanleykywu/federated-autoencoders

Repository files navigation

Read our paper here

In order to run our experiments please see client.py for CLI implementation details.

The help from our CLI tool shows the following:

git:(main) -> python client.py --help
usage: client.py [-h] [--dataset {fmnist,gtsrb,cifar10}] [--classes CLASSES [CLASSES ...]] [--epochs [EPOCHS]] [--latent_size [LATENT_SIZE]] [--verbose] [--type {client,server}]

optional arguments:
  -h, --help            show this help message and exit
  --dataset {fmnist,gtsrb,cifar10}
  --classes CLASSES [CLASSES ...], --names-list CLASSES [CLASSES ...]
  --epochs [EPOCHS]
  --latent_size [LATENT_SIZE]
  --verbose
  --type {client,server}

Please note that you will need to update the number of clients in the client.py at the Line 237 based on the number of clients.

Please see our files in log/ for examples of usage with multiple clients, one for each class and the expected outputs.

About

Use federated learning to train variational auto-encoders on disjoint distributions

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages