Understanding the Theory Behind Diffusion Models
This chapter will dive into the theory that powers diffusion models and see the internal workings of the system. How could a neural network model generate such realistic images? Curious minds would like to lift the cover and see the internal workings.
We are going to touch on the foundation of the diffusion model, aiming to figure out how it works internally and pave the foundation to implement a workable pipeline in the next chapter.
By comprehending the intricacies of diffusion models, we not only enhance our understanding of the advanced Stable Diffusion (also known as latent diffusion models (LDMs)) but also gain the ability to navigate the source code of the Diffusers package more effectively.
This knowledge will enable us to extend the package’s features in line with emerging requirements.
Specifically, we will go through the following topics:
- Understanding the image-to-noise process
- A more efficient...
