This project was submitted by Hong Jeon. It was one of the top submissions in our AI Alignment course (Mar 2024). Participants worked on these projects for 4 weeks.
This Colab was created to help you (and me) understand the inner workings of an image diffusion model.
After going through this Colab you will:
Learn the basic theory behind a denoising diffusion model
Implement a toy MLP model from scratch to generate MNIST images
Connect the learning from the toy model to real models, such as Stable Diffusion
Read the full piece here.