Does anyone have experience training diffusion models on custom data set? For example, I have a large corpus of domain specific 2D images and I would like to train a model like stable diffusion to generate novel new images that are plausible in that domain set?

Following as I am currently looking into the same area. Was just reading: https://stablediffusionapi.com/docs/dreambooth-finetunning/overview (If that works)

thanks @andres santanilla!