stablediffusion
stablediffusion copied to clipboard
depth2img mode with mask?
Is it possible to use depth2img with an image mask (i.e., in-painting)?
I'm trying to work my way through the scripts themselves and am still trying to grok what exactly is going on in text2img, img2img, and depth2img. What is the best resource for understanding the architecture of StableDiffusion, particularly as-implemented?