search

MIC-DKFZ / basic_unet_example

An example project of how to use a U-Net for segmentation on medical images with PyTorch.
Apache License 2.0
140 stars 38 forks source link

CPU usage is too high during training #9

Closed Zmjcc closed 4 years ago

Zmjcc commented 4 years ago

When I was training with another dataset, I set the patch to 512. During the training, the CPU occupation rate was too high, but during the verification, the occupation rate was not that high. Can you give me some suggestions on how to reduce the CPU occupation rate during training? Thanks!

elpequeno commented 4 years ago

Hi zmjcc, sorry for the delay, I was out of office. What do you mean by "too high". How is the CPU occupation on your own dataset different compared to the example dataset? Are you training on CPU or GPU?

elpequeno commented 4 years ago

Any update on this one?

elpequeno commented 4 years ago

As I am not 100% sure what the problem is and if it is still there, I will close this issue. If this problem still exists, please reopen it.