Prompts are embedded in the PNGs. I used the Dynamic Prompts and Regional Prompter extensions, and a wildcard setup that didn’t embed because of all the variables. This managed to keep the artists mostly separate from their art.
Models were CyberRealistic 3.1 and/or qgoPromptingReal (which did much better at understanding the prompt).
A bunch of these are slightly different but my basis was:
${age=25yo} ${person_description=!with perky breasts and ({short pixie-cut|shoulder-length|tied up}) __haircolor__ hair} ${context=against a stark empty wall } ${location=at an avant garde art gallery show} ${perspective=orthographic} An ${perspective} photo of a ${age} artist beside a self-portrait ${location}. ADDBASE
An ${perspective} photo of an embarassed ${age} artist ${person_description} wearing a contrasting dress and smiling at the viewer ${location}. ADDCOL
An ${perspective} photo of a single life-size beautifully expressive pornographic self-portrait of an identical nude ${person_description} {shocked|overwhelmed} by an (extremely intense orgasm:1.2) while (riding:1.1) a (painting of a pussy vibrator:1.2), gasping, moaning, (disheveled:1.2), ${context} ${location}.
After you comment, I decided to give it a try. And I am getting really creepy results 😂
Finally installed the software on my PC. But even replicating your prompts I always get the artist nude. Do you know what I am missing?
I think the problem here is that my prompt is written to use syntax that AI Horde workers don’t understand.
The stuff with the variables is https://github.com/adieyal/sd-dynamic-prompts but you can work around that by just manually writing out the whole prompt.
The stuff with ADDBASE and ADDCOL (and the ability to set the ratio between columns, which I thi k was at like 2 to 3) is provided by https://github.com/hako-mikan/sd-webui-regional-prompter and I don’t think there’s a good workaround to do that on AI Horde. Someone would need to add support for different prompts in different image regions to https://github.com/Haidra-Org/AI-Horde-Worker and to the web front-end.
Otherwise the model sees just what you typed in the box across the whole image and doesn’t get any help putting some elements on the left and others on the right.
I finally installed in my local, I was missing extensions. Now works great! You bite me with this. I have not so much time these days, but this is really funny to experiment and create stuff. So for sure you will see some posts from me in a future. For those interested, I am running it on a Ryzen 5 CPU and it takes forever depending on the config, but is feasable (an image in 15/20 min¿?). Would be great if moderators create some kind of tematic contest periodically. Something like “Sunbathing on the beach”, “Fun in the garage” . Just throwing the idea to keep this more active
Some of the best :)
Contests around a theme are a great idea. I think that’s the main power of this technology. Anybody can be naked, but only with AI can you have people be naked on the surface of the sun! Because in real life you would instantly die!