kewellcjj / pytorch-multiple-style-transfer Goto Github PK

Hi, can you please tell why you are using such high values for content-weight and style-weight arguments here?

Hi,
I am not very experienced in this, but I thought this was incorporating multiple styles in a single image? In your eval command, we are only providing one id for style - can we provide multiple ids there, for multiple styles?
Thanks!

can you share your dataset and trained model please?
ths!

I was running the training script but got:
Traceback (most recent call last):
File "neural_style.py", line 231, in
main()
File "neural_style.py", line 225, in main
train(args)
File "neural_style.py", line 70, in train
style = torch.stack(style_batch).to(device)
RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 0. Got 3 and 4 in dimension 1 at /pytorch/aten/src/TH/generic/THTensor.cpp:711
any ideas?

In the original TF implementation the instancenorm layers in the residual blocks are also conditioned as far as I understand.

Leaving them unconditioned in current implementation may lead to lesser quality of the stylization.