Generative AI in Practice: Transformers and Diffusion Models

The book is a comprehensive practical guide to modern generative artificial intelligence. It sequentially guides the reader from the basics of information representation to advanced methods of creating images, text, and audio using open models. Key architectures are thoroughly discussed: transformers and autoencoders, CLIP, diffusion models, and Stable Diffusion. A significant part of the book is dedicated to transfer learning, including fine-tuning language models and models for image generation. Creative applications, audio generation, and rapidly evolving directions in this field are detailed. The book has a pronounced practical focus, containing step-by-step instructions, projects for independent work, exercises, and tasks for consolidating the material.