-   Notifications  You must be signed in to change notification settings 
- Fork 9.8k
Closed
Description
Training a model for fast-neural-style raises a RuntimeError from variable division during input normalization.
- python2.7
- torch==0.1.12.post2
- torchvision==0.1.8
Traceback (most recent call last): File "neural_style/neural_style.py", line 226, in <module> main() File "neural_style/neural_style.py", line 220, in main train(args) File "neural_style/neural_style.py", line 65, in train style_v = utils.normalize_batch(style_v) File "/home/paperspace/embro/neural_style/utils.py", line 42, in normalize_batch batch /= Variable(std) File "/usr/local/lib/python2.7/dist-packages/torch/autograd/variable.py", line 793, in __idiv__ return self.div_(other) File "/usr/local/lib/python2.7/dist-packages/torch/autograd/variable.py", line 323, in div_ raise RuntimeError("div_ only supports scalar multiplication") Metadata
Metadata
Assignees
Labels
No labels