In the last chapter, we used a long short-term memory (LSTM) model on a time-series forecasting task. In this chapter, we will create a generator model, which means the model will not output predictions but rather files (in this case, images). We created a generator model in Chapter 7, Deep Learning for Natural Language Processing; however, in that case, we just generated latent features. Here, we will describe the main components and applications of generative adversarial networks (GANs). You will learn about the common applications of GANs and how to build a face generation model using a GAN.
Over the course of this chapter, we will investigate the architecture of a GAN. A GAN is composed of two competing neural networks, one of which is known as the generator model. It takes random data and creates synthetic target data. The other...
