This video explain common methods to generate images such as Autoencoders and Generative Adversarial Networks
Part 4: Transformers
Transformers have taken over natural language processing and show a lot of potential in computer vision. This video gives an overview of how they work.