chapter 0

0 Introduction

 
“I knew there was a way to blend records together, but I didn't know how to. This was haunting me when I was in my teens. In my frustration, I decided to start experimenting with electronics. I tested the torque factor on different turntables. I had to figure needles out.”

- Grandmaster Flash

This chapter covers

  • Understanding the basics of Generative AI
  • The benefits of using open models and open source tools for image generation
  • Details of hardware requirements for reproducing the results in this book
  • See what Stable Diffusion can do and get a sense of the skills you’ll learn in this book

You likely didn’t expect a book on Stable Diffusion to begin with a quote from one of the founding fathers of hiphop, but there’s an illustrative parallel to be made: Early hiphop mixed records in a similar way that Stable Diffusion mixes images; taking the familiar and rearranging it in a way that it becomes something never seen before.

All of us, since we were children bored in class, have doodled in notebooks hoping to transmit some image from our mind to the listless lined paper in front of us. Usually without the success we were hoping for. Generative AI allows us to revisit those childhood aspirations with renewed vigor, and far more impressive results!

What are “Generative Models?”

What are “models”?

Generative vs discriminative models

Open Models and Open Source Tools

Open model vs Open Source

Open Source tools we’ll be using

Hardware for Running Stable Diffusion

Running models with a GPU

Working with Unified Memory on Apple computers

Quick Glimpse of Stable Diffusion in Action

Where we’re going!

Creating “Art” with AI?!

Summary