tensorflow implementation of "Editable GAN" ( 128 by 128 )
Editable Generative Adversarial Networks: Generating and Editing Faces Simultaneously

• Requirement

  • tensorflow 1.8
  • scipy
  • tensorpack for MNIST
  • tqdm
  • python 3.6

• Files

  • compare_ops
    • ops.py from https://github.com/google/compare_gan (old version of compare_gan)
  • mnist_main.py / mnist_model.py
    • experiment on mnist with rotation or color
  • main.py / model.py
    • experiment on celebA
  • data_loader.py
    • Dataset implementation for "celebA".
    • Read image and label - label can be determined ( change "atts" in [a-z]{2-4}gan_celebA not "att_dict")
    • three modes - val : 10,000 / test : 10,000 / train : others
  • utils.py
    • Just utils ...
  • ops.py
    • Just ops ...

• DATASET ( for celebA )

  • for example,

    data
     |--- img_blah_blah ( ex img_celeba_128_center )
     |--- img_align_celeba
                 | ---- 000001.png ( or jpg.. change the code )
                 | ---- 000002.png ...
     |--- list_attr_celeba.txt
  

  • Then, execute : python main.py --gpu 1 --img_dir ../data/img_align_celeba --txt_dir ../data/list_attr_celeba.txt