Part 2 Introduction to
diffusion models
Diffusion models have rapidly emerged as one of the most powerful techniques for generative tasks, especially in image generation. We walk you through the core idea in chapter 5: generating an image by reversing a noise-adding process. You’ll implement a diffusion model step-by-step, starting from the basics of forward and reverse diffusion processes.
Once the fundamentals are clear, we move into conditioning and scaling. In chapter 6, you’ll learn how to guide diffusion models with conditioning information so that the outputs aren’t random but aligned with your intent. Chapter 7 tackles the challenge of generating high-resolution images and exploring techniques that allow diffusion models to scale up in quality and fidelity. These chapters give you the conceptual and coding foundation for understanding how today’s state-of-the-art text-to-image models are built.