search

stefanknegt / Probabilistic-Unet-Pytorch

A Probabilistic U-Net for segmentation of ambiguous images implemented in PyTorch
Apache License 2.0
270 stars 67 forks source link

Segmentation Fault (Core dumped) #2

Closed justusschock closed 5 years ago

justusschock commented 5 years ago

Hi, after trying to use your code, i got the following messages:

Loading file output_part2.pickle
Loading file output_part3.pickle
Loading file output_part5.pickle
Loading file output_part4.pickle
Loading file output_part1.pickle
Number of training/test patches: (13602, 1511)
No of classes: 1
No of input channels: 1
No of filters first layer: 32
Padding: 1
Segmentation fault (core dumped)

The training script is run with default parameters, I used recent master of your repo and pytorch (and added the custom kl divergence) and the Segfault seems to happen, during model initialization.

Do you have any ideas on how this could happen?

justusschock commented 5 years ago

Sorry, just noticed that this issue is due to my custom compilation being with CUDA but without cudnn

stefanknegt commented 5 years ago

Thanks for trying the code! Happy to hear the issue is resolved.