Communication-Efficient Adaptive Batch Size Strategies for Distributed Local Gradient Methods

Open in new window