Diffusion, Flow Matching, and Visual Generation
6.23.1DDPM, DDIM, and Denoising Diffusion#
6.23.2Score-Based Modeling and the Sampling Process#
6.23.3Latent Diffusion and VAEs#
6.23.4U-Net, Diffusion Transformer (DiT), and MMDiT#
6.23.5Flow Matching and Rectified Flow#
6.23.6Classifier-Free Guidance and Conditional Generation#
6.23.7Distillation, Consistency Models, and Few-Step Generation#
6.23.8Video Diffusion, Spatiotemporal Attention, and Long-Video Generation#