Stanford CME296 Diffusion and Large Vision Models | Spring 2026 | Lecture 5 - Architectures
from Stanford Online: University Courses & Lectures Hosted by Stanford Online
Learn more details about this course: https://online.stanford.edu/courses/cme296-diffusion-and-large-vision-modelsTo follow along with the course schedule and syllabus, visit: https://cme296.stanford.edu/syllabus/Chapters:00:00:00 Introduction00:05:26 Objective00:09:58 Convolutions, filters00:14:44 Receptive field00:17:14 Pooling00:19:06 U-Net00:27:52 Timestep representation00:30:31 Class label representation00:33:21 Timeline of U-Net models00:35:43 Diffusion Transformer (DiT)00:48:08 Adaptive layer normalization (adaLN)01:02:30 DiT end-to-end example01:12:57 Multimodal DiT (MM-DiT)01:23:33 Qwen-Image, Z-Image, FLUX.101:24:27 Timeline of DiT models01:25:25 Absolute position embeddings01:38:48 Rotary position embeddings (RoPE)01:39:59 2D RoPE variantsUploaded 11-05-26For more information about Stanford’s graduate programs, visit: https://online.stanford.edu/graduate-educationAfshine Amidi is an Adjunct Lecturer at Stanford University.Shervine Amidi is an Adjunct Lecturer at Stanford University.View the course playlist: https://www.youtube.com/playlist?list=PLoROMvodv4rNdy8rt2rZ4T2xM0OjADnfu
Embed this episode
No transcript yet.