contrastive-unpaired-translation
contrastive-unpaired-translation copied to clipboard
taesungp
Contrastive unpaired image-to-image translation, faster and lighter training than cyclegan (ECCV 2020, in PyTorch)
In Eq 3 of original paper, it seems that $\hat z^s$ and $z^s$ are positive pairs. It is reasonable in the 'horse' and 'zebra' examples because the location of object...
Thanks for releasing the code! Could you release the GTA->Cityscapes pretrained model? Have you ever tested the fid of this results?
`self.nce_layers` is initialized to `[0,2,4]` by default. While when using `smallgan2` as generator there is no any feature coming from 4th layer. The patches are only coming from `[0,2]` in...
When I use the "--model test" option to run a trained model on new input I get an error: No such file or directory: './checkpoints/apple2orange_DCL/latest_net_G_A.pth' The checkpoint directory has only...
Hello, I am trying to run test.py script but got this error. Can you help me to fix this? Thank you! RuntimeError: Error(s) in loading state_dict for ResnetGenerator: Missing key(s)...
I want to ask that why my generator seems to generate a same output picture in some epoch. The loss function remains a relatively stable value? Is it a mode...
Steps to reproduce the issue: ``` git clone https://github.com/taesungp/contrastive-unpaired-translation CUT cd CUT conda env create -f environment.yml conda activate contrastive-unpaired-translation bash ./datasets/download_cut_dataset.sh grumpifycat python train.py --dataroot ./datasets/grumpifycat --name grumpycat_CUT --CUT_mode...
The CUT paper and this repo claim the discriminator is a 70x70 PatchGAN. However, I verified that this is not the case. Searching the repo for "70" only yields the...
the code produces error with grayscale images. My input crop size is 400, and the number of input and output channel is set to 1, then I get this error:...
