Mainly for realistic and detailed models.
Compilation of 100 images from my "fantasy" works.
I knew it would be a difficult style to replicate, the result is correct but not perfect for realistic models, for anime models, the style is variable but it can give interesting results.
The images will probably need to be reworked.
Strength: 0.7 -> 1
You can add: (Frank Frazetta art style) or (Frank Frazetta painting style), the AI knows this name and can adjust the style (realistic models), also "realistic illustration", "realistic screencap" and (detailed matte painting:1.2).
Illustrious version (tests not finished yet):
Favor realistic models (even if I haven't found a good one for this LoRa yet) or semi-realistic ones.
Setting: 0.7/0.8 (avoid going higher)
