Distributed Deep Neural Network Training with Important Gradient Filtering, Delayed Update and Static Filtering.

Autor: Li, Kairu, Wu, Yongyu, Tian, Jia, Tian, Wentao, Ye, Zuochang
Zdroj: ACM International Conference Proceeding Series; 12/20/2019, p120-125, 6p
Databáze: Complementary Index