This is the paper on which our work for Realistic Image Generation is based. For realistic image generation we used GAN with differentiable augmentation on generator and discriminator.
Repository on which is our code based is here.
We trained GAN on CIFAR-10 dataset, Obama-100 dataset and also our work is to train GAN on animal dataset
Here you can see Colab files for image generation of 4000 animal images for every class (cats, dog and wild animals) from random chosen 200 images for every class in dataset.
Also there are Colab files for image classification using Alexnet for original and augmented dataset.