Repository for Pytorch Implementation of Least Squares Generative Adversarial Networks
Regular GANs hypothesize the discriminator as a classifier with the sigmoid cross entropy loss function.
This loss function, however, may lead to the vanishing gradient problem during the learning process.
LSGANs(Least Squares GAN) adopt the least squares loss function for the discriminator.
The paper shows that the objective function for LSGAN yields minimizing the Pearson chi-square divergence.
