=================
Diffusion serving
=================
Diffusion models are a class of generative models that have gained popularity for their ability to generate high-quality images. They work by gradually adding noise to an image and then learning to reverse this process, effectively denoising the image step by step.

Milestone models
-----------------

Stable Diffusion

Stable video Diffusion

ControlNet

Diffusion transformer

Main challenges
-----------------
Computation burdens

Heterogeneous hardware

Memory demands


Popular optimizations
----------------------
xxx

Inference engines
-------------------
xxx