search

cvlab-yonsei / EWGS

An official implementation of "Network Quantization with Element-wise Gradient Scaling" (CVPR 2021) in PyTorch.
GNU General Public License v3.0
88 stars 16 forks source link

training time on ImageNet #3

Open qiulinzhang opened 3 years ago

qiulinzhang commented 3 years ago

Thanks for your great paper, can you share your total training time and the number of GPU used, or to say, GPU hours? because in my experiment of EWGS, it is very slow.

junghyup-lee commented 2 years ago

Using 4 GPUs (RTX 2080ti), training ResNet-18 on ImageNet roughly takes 60 hours. If the training time is far slower than this, you could try adjusting num_workers in a dataloader, since a proper number of workers may differ depending on the CPU spec.