search

bfortuner / pytorch_tiramisu

FC-DenseNet in PyTorch for Semantic Segmentation
MIT License
306 stars 66 forks source link

"Out of memory" on 6GB GTX1060 #9

Open jbmaxwell opened 6 years ago

jbmaxwell commented 6 years ago

I tried batch_size=1, but no difference. I haven't attempted to calculate the actual memory requirement, but would you expect this to run on a 6GB GPU?

jbmaxwell commented 6 years ago

Oh, okay, seems like a false alarm—rebooting my ubuntu box seems to have solved the problem (also learned about clearing GPU memory in the meantime)!

Yaozuwu commented 6 years ago

Could you tell me your config? My config: cuda:9.0 pytorch:0.4.0 torchvision:0.2.1 but there are some warning and bug in my program, I have also get"Out of memory" Thank you very much

jbmaxwell commented 6 years ago

I'm actually still on cuda 8.0.61, with pytorch 0.3.1—I haven't run this for a while. But for me the solution was just a reboot and then adjusting the batch_size to the max possible before the memory error. (I think I also reduced my input size, though I realize that may not be possible in your case.)