Change To Pytorch Version: 1.3+ #20
Conversation
`torch.nn.Parameter(d.get(41).weight.float())` Change to: `torch.nn.Parameter(torch.Tensor(d.modules[41].weight))`
`load_lua()` change to: `load()`
```python
cImg = Variable(contentImg, volatile=True)
sImg = Variable(styleImg, volatile=True)
```
Change to:
```python
with torch.no_grad():
cImg = Variable(contentImg)
sImg = Variable(styleImg)
```
Add an URL to download the model while one in China.
There was a problem hiding this comment.
Causing one error now
$ python3 WCT.py --cuda
Transferring in1.jpg
Traceback (most recent call last):
File "/Universal Style Transfer/PytorchWCT/WCT.py", line 110, in
styleTransfer(cImg,sImg,imname,csF)
File "/Universal Style Transfer/PytorchWCT/WCT.py", line 56, in styleTransfer
Im5 = wct.d5(csF5)
File "/home/anaconda3/envs/yolo/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/Universal Style Transfer/PytorchWCT/modelsNIPS.py", line 885, in forward
out = self.reflecPad15(x)
File "/home/anaconda3/envs/yolo/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/anaconda3/envs/yolo/lib/python3.10/site-packages/torch/nn/modules/padding.py", line 178, in forward
return F.pad(input, self.padding, 'reflect')
RuntimeError: Padding length too large
Make the program runs well in Windows and Pytorch 1.3+.