Summary

You have come a long way from being introduced to one of the most promising areas in deep learning. Let's revisit some of the concepts that we learned about in this chapter.

We started this chapter by understanding what GANs are and their major applications. We then went on to understand the various building blocks of GANs, such as the real datasets, fake datasets, the discriminator operation, the generator operation, and the GAN operation.

We executed a problem statement to progressively build a fully connected GAN (FCGAN) to solve a real function. In the process of building the GAN, we also implemented exercises for creating real datasets, creating fake datasets, creating a generator network, creating a discriminator network, and finally combining all these individual components to create the GAN. We visualized the different plots and understood how the generated data distribution mimicked the real data distribution.

In the next section, we understood the...