A Gentle Introduction to Diffusion Models and Flow Matching
Diffusion models have emerged as a powerful and flexible class of generative models, underpinning recent breakthroughs in image, audio, and scientific data generation. This article offers a gentle, beginner-friendly overview of diffusion models and the related flow matching framework, demystifying their mathematical foundations and practical training procedures. We will delve into the core concepts, guiding equations, and step-by-step training algorithms, aiming to provide readers with both an intuition for how these models work and a roadmap for implementing them in practice. This article serves as notes on papers [1] and [2]. ...