search

nv-tlabs / GSCNN

Gated-Shape CNN for Semantic Segmentation (ICCV 2019)
https://nv-tlabs.github.io/GSCNN/
Other
915 stars 200 forks source link

some problems about training #14

Closed buzhiqimeiliuqiangdong closed 4 years ago

buzhiqimeiliuqiangdong commented 4 years ago

When I run 'python train.py --evaluate --snapshot checkpoints/best_cityscapes_checkpoint.pth',CUDA out of memory. image My GPU is 11GB of memory,How can I solve this problem，thx.

shubhaminnani commented 4 years ago

Hi, while I was training on 16GB GPU there was no problem. Problem is 12GB GPU. Thank you!

shoutOutYangJie commented 4 years ago

When I run 'python train.py --evaluate --snapshot checkpoints/best_cityscapes_checkpoint.pth',CUDA out of memory. image My GPU is 11GB of memory,How can I solve this problem，thx.

how about the size of your image

tovacinni commented 4 years ago

Out of memory issues are almost always caused by not having enough GPU memory. You could either use a smaller backbone network, or try to optimize the code by freeing up some variables as soon as they are done being used.