4 citations · 4 across the 1 of their papers we have counts for
1 paper
Mher Safaryan, Peter Richtárik
Various gradient compression schemes have been proposed to mitigate the communication cost in distributed training of large scale machine learning models. Sign-based methods, such…