Minibatch vs Local SGD for Heterogeneous Distributed Learning

Open in new window