Hello! Please check out my stable diffusion webui at Sdpipe Webui - a Hugging Face Space by lint, I would really appreciate your time giving it a try and any feedback!
Right now it supports txt2img, img2img, inpainting and textual inversion for several popular SD models on Huggingface.
It’s a lightweight implementation of the diffusers pipelines framework. Model weights are kept in memory when switching between pipelines/treaining textual inversion for fast response times. The app is simple enough that you can easily adapt it to your own use case.