Stanford CME296 Diffusion & Large Vision Models | Spring 2026 | Lecture 2 - Score matching

Stanford Online · 108:48

This lecture introduces score matching as a second generative-modeling paradigm alongside last week's DDPM: instead of maximizing a variational likelihood bound, you learn the score \(\nablax \log p(x)\) and use it as...

Read the full summary on tuber

Redirecting...