Pure tensorflow implementation of DCGAN
Sources:
- Improved Techniques for Training GANs
- Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
- GANHacks
- tensorflow 1.0
- h5py
- matplotlib
- opencv 3
- numpy
- tqdm
- colorama
- terminaltables
- parmap
CelebA data:
- Go to http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html
- In the Downloads section, select Align&Cropped images.
- In the dropbox page that follows, download the Anno, Eval and Img folders.
- Copy these folders to
/data/raw
. - Extract the zip files.
Follow these instructions