chainer implementation of triple GANs
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Sample code for Triple GAN


This software is a python implementation of Triple GAN using chainer.

See the paper.

Notice that this code had not implement gaussian noise in discriminator, Because I don't know how to implement it in chainer.


I only confirm the operation under this environment

  1. python 3.5.3
  2. chainer 3.3.0
  3. pillow 5.0.0
  4. numpy 1.14.0
  5. cupy 2.3.0


First of all, prepare mnist data as npy form.
python Then, you can get mnist_train_img.npy, mnist_train_label.npy, mnist_test_img.npy and mnist_test_label.npy.

To train this model, do python

Sample image

Sample image generated in generater after about 500 epochs. resultimage_580


I also made triple GAN code in tensorflow.
But, there are some bugs.