Part 2: Transfer Learning for Generative Models

In Part 1, we met the core building blocks behind many of the most popular generative models. What we did NOT do, apart from a small diffusion model demo, is train any of these models for ourselves. Training advanced models from scratch is a difficult task, and often requires a lot of data and compute. So, instead, we will be using a technique called transfer learning, which allows us to take a model that has been trained on a large dataset, and adapt it to a new dataset with much less data and compute.

Chapter 5 focuses on transformers and text, exploring… TODO

Chapter 6 focuses on image generation, exploring… TODO

Chapter 7 covers some newer methods for more efficient fine-tuning… TODO

Together, these chapters will give you a solid foundation for training your custom generative models…