The Algos (AI Engines) used on NightCafe have a VERY limited training-image background to work with. And also very limited cultural context. If you want something out of their programming, or something more specific and detailed then they can focus on in a few seconds of runtime, then you will need to upload more focused images, and evolve them. Like photographs of animals or plants that they don't 'know' about, haven't been trained on, or mythological figures from cultures they are not trained on. In my case, as an astrophysicist, I have had to upload images of (Helical, Toroidal) Magnetic Fields which they do not understand. They don't know what a Star Constellation is, so you cannot ask them to draw Scorpio, or the Southern Cross, or other. So I upload photographs taken by Telescopes and Space Telescopes, etc. and use the prompts to add iridescent backgrounds, to zoom in on the images, etc.
a redhair nice happy Autumn female holding pumpkin in gentile hands. The girl should be depicted with small, detailed embellishments that stand out against the contrasting dark green velvet of her dress. The frame of the illustration features Irish pattern elements.
Weight: 1
heavenly sunshine beams divine bright soft focus holy in the clouds