Effects of Dataset properties on the training of GANs

Ilya Kamenshchikov, Matthias Krauledat

Generative Adversarial Networks are a new family of generative models, frequently used for generating photorealistic images. The theory promises for the GAN to eventually reach an…