HI,I want use the the dcgan from pytorch examples to generate faces, and i want to use training and testing images that I collected by myself. So how can I create dataset like cifar-10-batches-py? Thanks very much!
you can use the ImageFolder
dataset for this purpose: https://github.com/pytorch/vision/#imagefolder
You are right, the main.py file in pytorch examples of dcgan can be used for the task of collecting my owen images for generating fake images.
Thanks very much!