MDGAN — Simple TensorFlow Implementation [Paper]
fake_img = generator(noise)
real_logit = discriminator(real_img)
fake_logit = discriminator(fake_img)
g_loss = generator_loss(fake_logit)
d_loss = discriminator_loss(real_logit, fake_logit)
Simple Tensorflow implementation of "MDGAN: Mixture Density Generative Adversarial Networks" (CVPR 2019)
PythonMIT
fake_img = generator(noise)
real_logit = discriminator(real_img)
fake_logit = discriminator(fake_img)
g_loss = generator_loss(fake_logit)
d_loss = discriminator_loss(real_logit, fake_logit)