search

eriklindernoren / PyTorch-GAN

PyTorch implementations of Generative Adversarial Networks.
MIT License
16.22k stars 4.05k forks source link

Don't know how to produce datasets for cycleGAN. #143

Closed zherlock030 closed 3 years ago

zherlock030 commented 3 years ago

Hey Im wondering how to produce my on datasets for cycleGAN. In cyclegan.py I see real_A = Variable(batch["A"].type(Tensor)) real_B = Variable(batch["B"].type(Tensor)) I want to know where we define the batch['A'], is it because of the foldername and defalutly defined in dataloader? I want to know where we define this mechanism since the input data belong to different classes and I want to distinguish it. THX.

zherlock030 commented 3 years ago

ohhh I find it in datasets.py