// ✦ Diffusion, Flow Matching & Generative Models
Latent Diffusion Models
Runs diffusion in a VAE's compressed latent space to cut cost.
潜空间扩散（Stable Diffusion） 2021 Intermediate core
IN PLAIN WORDS
Pixel diffusion is expensive. Compress images ~8× with an autoencoder, denoise in latent space and inject text via cross-attention — the core of Stable Diffusion.
Used in: Stable Diffusion family
TIMELINE · prerequisites → this → what builds on it
01Original paper
High-Resolution Image Synthesis with Latent Diffusion Models
Robin Rombach, Andreas Blattmann, Dominik Lorenz, et al. · 2021
arXiv ↗ ✓ Verified · arXiv API
02Where it sits
Learn the left column first; the right column builds on this one.
Learn first
Then you can learn
Related:Classifier-Free Guidance (CFG)
Left: prerequisites · right: builds on this · dashed: related. Full map →
03Learning resources
Official docs, explainers, courses and reference code. All links checked.
Explainers
Reference code
Official docs
04Where the course teaches it
On learning paths:Generative Models & Frontier Research
For AI / data
- Verification: verified, checked 2026-10-21.
- JSON: /api/learn/algorithms.json (id:
latent-diffusion)