Art-generating AIs are fantastic and developments are happening at a breakneck pace. While experimenting with Stable Diffusion (via the fantastic Automatic1111 Webgui) to generate images I made a document to keep track of what I learned and the ways I was fixing problems.
Here is my HOW TO guide which includes my recommended settings and techniques for:
- How to use inpainting to redraw a small messed-up part in a generated image
- How to use inpainting to remove parts of an image
- Use inpainting to redraw areas or details
- Generate same-but-different variations on an image
- And more!