CMU 10799 S26: Lecture 2 - Denoising Diffusion Models - Diffusion & Flow Matching

AI • Machine Learning • Tech on Instagram: "Diffusion models such as Denoising Diffusion Probabilistic Models (DDPMs) work by first adding noise to images iteratively until the image becomes pure noise. The model learns how this noising process works by predicting the noise added at each step. Then, the process is "reversed"; the model starts from random noise and iteratively predicts and subtracts the entire noise portion of the image. This way, small amounts of noise are removed, refining the