that's damn impressive, my man. it'd probably have been a lot easier to use something else to generate a base image, then the image to image AI generation with the proper prompts to fine tune them.
