发帖
雷达卡
Momentum Residual Neural Networks
Michael E. Sander 1 2 Pierre Ablin 1 2 Mathieu Blondel 3 Gabriel Peyre 1 2
Abstract 2019). Most of deep learning tasks involve graphics process-
ing units (GPUs), where memory is a practical bottleneck in
The training of deep residual neural networks
several situations (Wang et al., 2018; Peng et al., 2017; Zhu
(ResNets) with backpropagation has a memor ...
京公网安备 11010802022788号
