MiCRO: Near-Zero Cost Gradient Sparsification for Scaling and Accelerating Distributed DNN Training

Re-assign community