Artemis: tight convergence guarantees for bidirectional compression in Federated Learning

Open in new window