The CelebA dataset (https://www.kaggle.com/datasets/jessicali9530/celeba-dataset) was used to generate images with two different approaches, using Variational Autoencoders (VAEs) and using Generative Adversarial Networks (GANs). This was just an exercise so results are not good. For to achieve good results, much more iterations and time working on the architectures are needed.
