ScaleNet icon indicating copy to clipboard operation
ScaleNet copied to clipboard

Training time

Open lxtGH opened this issue 6 years ago • 1 comments

Hi! Would you share your training time and specific hardware enviroment for trainning your network from scratch?

lxtGH avatar Jun 01 '19 12:06 lxtGH

Training time varies from different depth. Multi-branch in Pytorch can not be processed in parallel, the training time is little slower than ResNets. TensorFlow runs faster.

Eli-YiLi avatar Jun 24 '19 03:06 Eli-YiLi