Table of Contents
Diffusion models are inspired by non-equilibrium thermodynamics. They define a Markov chain of diffusion steps to slowly add random noise to data and then learn to reverse the diffusion process to construct desired data samples from the noise. Unlike VAE or flow models, diffusion models are learned with a fixed procedure and the latent variable has high dimensionality (same as the original data).– Lil’Log (source)
Please visit the source above to get more details on this or check out this video.
I don’t want to get caught up on the exact details of various papers since I don’t want to misquote anything. I want to cite that the amazing group of people got together and used the various diffusion options along with CLIP to create Disco Diffusion🕺 which is a tool I have been using for months now to create amazing images and videos.
I do have to personally say this has been so far my favorite way to generate AI artwork. I started creating pieces of art with style transfer about 2 1/2 years ago and have watched this space quickly grow and am super excited about what we have now and what’s coming. Also a super big shoutout to everyone involved in creating, updating and maintaining this stuff.
As noted on their official GitHub page Disco Diffusion is defined as “A frankensteinian amalgamation of notebooks, models and techniques for the generation of AI Art and Animations.”
There are various contributors but the most notable are:
You can check out the rest of the contributors here.
Most of you who have stumbled onto this page have been using Disco Diffusion and looking for hints and tips or just getting started and you want a jumping off point to get you rolling!
List of Links I recommend to get you going or advance your understanding of DD
Here are a few of my DD Image generated So far.