The Limit of the Batch Size

Wang Yuhui
Wang Yuhui
Zhang Huan
Zhang Huan
Zhang Zhao
Zhang Zhao
Cited by: 0|Bibtex|Views32|Links

Abstract:

Large-batch training is an efficient approach for current distributed deep learning systems. It has enabled researchers to reduce the ImageNet/ResNet-50 training from 29 hours to around 1 minute. In this paper, we focus on studying the limit of the batch size. We think it may provide a guidance to AI supercomputer and algorithm designer...More

Code:

Data:

Your rating :
0

 

Tags
Comments