flâneur — a map of the web's best reading

⭐️ Diffusion Models

andrewkchan.dev · 6,310 words · saved by 3 readers

I spent 2022 learning to draw and was blindsided by the rise of AI art1 models like Stable Diffusion. Suddenly, the computer was a better artist than I could ever hope to be. It’s been two years, and image generation with diffusion is better than ever. It’s also led to breakthroughs in animation, video generation, 3D modeling, protein structure prediction, and even robot trajectory planning. Where did it come from, how does it work and where is it going? This post collects my notes on the theory of diffusion and applications to image generation and other tasks. Readers should know some probability theory (Bayes’ rule, Gaussian distributions). Examples and code using PyTorch are provided. The basic problem of generative modeling is: given a set of samples from an unknown distribution 𝑥 ∼ 𝑝 ( 𝑥 ) x∼p(x), we want to generate new samples from that distribution. Generative adversarial networks treat this as a game: a generator model taking a random seed is trained to fool a discriminato

Contents Diffusion Models Notes on the theory behind models like Stable Diffusion and their applications. Supplement: Discussion on Hacker News . I spent 2022 learning to draw and was blindsided by the rise of AI art There is lots to say about this, whether it has been or will be a good thing for artists and society in the long run. I hope to write about it in another post. models like Stable Diffusion. Suddenly, the computer was a better artist than I could ever hope to be. It's been two years, and image generation with diffusion is better than ever. It's also led to breakthroughs in animatio

Explore this link on the map →

saved by

related reading