ATOMO: Communication-efficient Learning via Atomic SparsificationPublished in NeurIPS, 2018Share on Twitter Facebook LinkedIn Previous Next