Open Assistant + Stable Diffusion

Сделал прототип Personal Art Assistant. Для этого я подал в Open Assistant промпт:

Examples of high quality prompt for portrait of a beautiful female wizard for text-to-image Ai image generation

– (extremely detailed CG unity 8k wallpaper),(masterpiece), (best quality), (realistic), cute anime old french village, in the mountains, blue sky, summer, flowers, plants, highly detailed, style of master anime, perfect perspective, intricate detail

Give me 4 more examples of the same scene, but in different times of day and different seasons and different clothing styles and accessoires.
Rules:

don't use any pronouns
avoid using these words: in a, the, with, of, the, an, and, is, by, of.
start every prompt with these exact words: " ((best quality)), ((masterpiece)), ((realistic)), ((beautiful young female wizard)), "
end every prompt with these exact words: " on eye level, scenic, masterpiece "
leave two empty lines after the end of each prompt

На выходе получил 4 промпта, которые подал в Stable Diffusion со следующими параметрами:

- script: Prompts from file or textbox
- model: revAnimated
- Sampler: DPM++ 2S a Kerras
- Steps: 20
- Resolution: 512x768
- Restore Faces
- Hires. fix: 1.25
- negative embeddings: verybadimagenegative, badhandv4
- negative prompt: kitsch, ugly, oversaturated, grain, low-res, Deformed, blurry, blur, out of focus, poorly drawn, childish, mangled, old, surreal, text, badhandv4, verybadimagenegative_v1.3, 2 heads, old

- - -
Artifactory.ai - запуск Stable Diffusion в облаке за 5 минут