I implemented the GAN on CIFAR dataset. Ran it for 10k epochs.
I/p image dimensions - (32 * 32 * 3) There are 60k images in CIFAR10 dataset and 10 classes. Latent dimension = (1 * 100) This is the noise that we give as input to the generator learning of Activation function (LeakyReLU) - 0.2 momentum - 0.8 Optimizer - Adam