Optimizing Pipelined Computation and Communication for Latency-Constrained Edge Learning

Open in new window