Hugging face diffusers and it allows one to customize the inage generation pipeline.
Diffusers
Overview
Huggingce Diffusers is the go-to library for state-of-the-art pretrained diffusion models for generating images, audio, and even 3D structures of molecules. It can be used to:
The simple inference solution
Training customize diffusion model
It focus on:
Usability over performance
Simple over easy
Customizability over abstractions
The main components
It has three main componenets:
State-of-the-art diffusion pipelines for interence with code
Interchangeable noise schedulers for balancing trade-offs between generation speed and quality
Pretrained models that can be used as building blocks, and combined with schedulers, for creating end-to-end diffusion systems