Description
This is a general breakdown of how I generate my images.
It often starts with many generations from a raw prompt, though this time I used controlnet to get near what I had in mind. Then After seeing one or 2 images that may work I begin editing them and painting in very very roughly to give Stable Diffusion an idea of the placement of limbs and features. This usually means generating at least another 5-30 images and photobashing them together.
This process repeats for all parts of the image with smaller and smaller regions then I finally upscale it.
givemeyourteeth
MemberThis is a good resource
Login to respond »