Learn what GAN is and get a high-level view of it with PyTorch.

Machine Learning with Python Libraries

This module introduces  generative adversarial networks (GANs) and a practical step-by-step tutorial to make our own GAN with PyTorch. 

Through this module, we’ll learn to build GANs with industry-standard tools. In the first section, we’ll dive into PyTorch and refresh our understanding of neural networks by building a simple image classifier. 

In the second section of this module, we’ll explore the idea of adversarial training and build more sophisticated GANs progressively. We’ll do this by first learning a simple 1010 pattern, then monochrome images of handwritten digits, and finally fully-colored images of faces. The third and last section extends the core GAN concept. We apply it to convolutional neural networks, and develop a conditional GAN to generate data of the desired class. 

By the end of this module, we’ll have a solid understanding of how to build GANs for our machine learning projects.

PyTorch (GAN Using PyTorch)

Reinforce your understanding and test your knowledge of the topics covered in this chapter.

Getting Started

Generative Adversarial Networks Fundamentals

Best Practices for Model Design and Training

Building Our First GAN with PyTorch

Generating Images Based on Label Information

Image-to-Image Translation and Its Applications

Image Restoration with GANs

Training GANs to Break Different Models

Image Generation from Description Text

Sequence Synthesis with GANs

Reconstructing 3D Models with GANs

Concluding Remarks

Appendix

Quiz: Generative Adversarial Network Fundamentals