READ THIS FIRST

This repo is an anonymized version of an existing repository of GitHub, for the AIStats 2021 submission: SparseFed: Mitigating Model Poisoning Attacks in Federated Learning with Sparsification. So if you see another repo that looks identical to this, we are not stealing anyone's code, that's my repo.

SparseFed

This repo contains an implementation of model poisoning attacks on a federated learning system.

It comes with a few experimental setups; various Residual Networks on CIFAR10, CIFAR100, FEMNIST, ImageNet (cv_train.py) and GPT2 on PersonaChat (gpt2_train.py) (attack is currently not implemented for PersonaChat)

There are a variety of command-line args which are best examined by looking at utils.py

The server is contained in fed_aggregator.py and the worker is contained in fed_worker.py

To use sketching, you need to install https://github.com/nikitaivkin/csh