Generative models are widely used in many subfields of AI and Machine Learning. Roughly speaking, there are four types of generative models in the deep learning field, including variational autoencoders (VAE), generative adversarial networks (GAN), autoregressive models, and normalizing flow models. In this talk, I will give a big picture of generative models in deep learning but focus on normalizing flow models, from the first normalizing flow model introduced in 2015, to the state-of-the-art models invented this year (2020). In the end, if time permits, I will also discuss some open problems in deep generative models and my solutions