Stanford CME296 Diffusion & Large Vision Models | Spring 2026 | Lecture 8 - Trending Topics

Short summary

Stanford's CME296 Lecture 8 covers diffusion and vision model architectures, training methods, state-of-the-art models, and applications in image/video generation and LLMs. A 110-minute technical deep dive for product teams building AI-powered features.

• •Comprehensive coverage of diffusion model theory, score matching, and flow matching techniques
• •Modern vision model architectures including guidance, training approaches, and evaluation metrics
• •Applications across image generation, video generation, and LLM integration

Generated with AI, which can make mistakes.