search

bigmb / Unet-Segmentation-Pytorch-Nest-of-Unets

Implementation of different kinds of Unet Models for Image Segmentation - Unet , RCNN-Unet, Attention Unet, RCNN-Attention Unet, Nested Unet
MIT License
1.87k stars 345 forks source link

question about torchsummary #13

Closed DLsnowman closed 4 years ago

DLsnowman commented 4 years ago

hello, in line 102:torchsummary.summary(model_test, input_size=(3, 128, 128)) I get a error: Error in dlopen or dlsym: libnvrtc.so.10.0: cannot open shared object file: No such file or directory. Is the problen about version of torchsummary?

bigmb commented 4 years ago

It can be due to its version. Which one are you using?

DLsnowman commented 4 years ago

Mine is torchsummary 1.5.1, pytorch 1.2.0. I can run: summary(torchvision.models.vgg11().cuda(), input_size=(3, 128, 128)). but can not summary() the U_Net instance whose class is written in Models.py file.

DLsnowman commented 4 years ago

When I use torch.cat() without cuda(): a = torch.tensor(np.random.random((3, 5))) b = torch.tensor(np.random.random((6, 5))) c = torch.cat((a, b), 0) There is no error. When I use cuda(): a = torch.tensor(np.random.random((3, 5))).cuda() b = torch.tensor(np.random.random((6, 5))).cuda() c = torch.cat((a, b), 0) The error happens: RuntimeError: Error in dlopen or dlsym: libnvrtc.so.10.0: cannot open shared object file: No such file or directory

DLsnowman commented 4 years ago

the problem is about the version of pytorch 1.2.0 I install pytorch1.1.0. The error disappears.

bigmb commented 4 years ago

Pytorch 1.1.0 just got stable, I suggest dont shift to 1.2.0 right now. I looked online into the issue. Couldn't find much on it.