confidencemanharoldhill commented at 2022-11-29 22:38:15 » #2767053

I'm currently working on my own stable diffusion style text to image system! The idea is to identify gelbooru tags in the structure of natural language, identify abstract objects/concepts from the relationships between them, then represent those as images.

Here's the prompt I wrote for this image's training data in case any of the rest of you want to use it for similar work:

Nino is a girl that dresses in the gyaru fashion style and has orange eyes and long blonde hair with bangs. She is standing alone and looking forward while licking her lips and squeezing her large breasts together. She is wearing earrings, a black choker with a heart on it, a pleated black skirt, and a white shirt that covers her erect nipples, but reveals her cleavage because she is wearing no bra. She is smiling with her mouth closed and tongue out, and is blushing. She has hair over one eye, but it is visible through her hair. The image is high res digital art in the style of sunaba suzume.

1 Points Flag