MIPD: An Adaptive Gradient Sparsification Framework for Distributed DNNs Training
Sign Up to like & getrecommendations! Published in 2022 at "IEEE Transactions on Parallel and Distributed Systems"
DOI: 10.1109/tpds.2022.3154387
Abstract: Asynchronous training is widely used for scaling DNN training over large-scale distributed deep learning systems using the parameter server architecture. Communication has been identified as the bottleneck, since large volumes of data are exchanged during… read more here.
Keywords: sparsification; mipd adaptive; framework; sparsification framework ... See more keywords