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

Daegun Yoon, Sangyoon Oh

Browse the full HiPC paper archive.