Lilian Weng’s long guide breaks down diffusion models from noise to samples
vboykis · x · 2026-07-28
A long Lilian Weng post explaining diffusion models in detail, from the forward noising process to reverse sampling.
It covers:
- the connection to stochastic gradient Langevin dynamics and noise-conditioned score networks
- classifier guidance and classifier-free guidance for conditioning
- ways to speed up sampling with fewer steps and distillation
- latent diffusion, higher-resolution generation, and model architecture choices
The article is a structured overview of how diffusion models work and why they became a major generative modeling approach.
More from Research
- Paper proposes an eight-part vocabulary for multi-agent research systems — Bardiya Akhbari · 2026-07-28
- A step-by-step hand derivation shows how residual connections power deep nets — ProfTomYeh · 2026-07-28
- AI agents should talk to each other, share context, and learn skills together — heyshrutimishra · 2026-07-28
- Graft argues code agents should inject repo context automatically, not wait for MCP calls — shhdwi · 2026-07-28
- Kimi is described with a linear-attention variant and attention residuals — burny_tech · 2026-07-28
- Nature Communications links NLP embeddings to flexible semantic retrieval in the brain — bttyeo · 2026-07-28