Greedy Low-Rank Gradient Compression for Distributed Learning with Convergence Guarantees

Open in new window