research
∙
11/29/2018
Data-parallel distributed training of very large models beyond GPU capacity
GPUs have limited memory and it is difficult to train wide and/or deep m...
share
research
∙
08/01/2016
