Crossover model between anime and semi-realism
Contours: clear, illustrative lines inspired by quill drawings
Shading & textures: painterly brushes with a subtle oil-painting character
Rendering: hybrid of 2D illustration and 3D depth for enhanced dimensionality
Focus: strong play of light and shadow, with atmospheric contrasts
Developed as an experiment, deliberately positioned in the space between two opposites. The goal was not to create a neutral midpoint, but to preserve the distinctive, complementary qualities of both sides at the same time.
I am happy if u anjoy and stay curious abt ur results ☺ best regards ~♥ Reij
♫~ if u like drawing ur visions with my model, a ♥ would make me smile ~♫
░▒▓ recommended setting examples:
832x1216 | 1040x1510
DPM++ 2M Karras / Euler a
CFG 4-7
+ PROMPT
masterpiece, best quality,- PROMPT
worst quality, bad quality, ░▒▓ my hires settings:
x 1,526 | Denoising strength 0,2 | Hires steps 30 | Hires CFG Scale 7
░▒▓ ADetailer for further distance recommended
Description
VAE baked in
merge reciepe: just my merged models ☺ and some special trained Loras ♥
FAQ
Comments (18)
Please some knows a ADetailer for further distance?
By that I meant: for very wide full body images that I would recommend using an ADetailer. depending on image settings and taste :)
@reijlita Sorry, I misunderstood, thanks for clarifying.
@colorgman001 everything is fine ☺ if you have any questions, feel free to ask them.
I've been working on it for a while now so that we don't necessarily have to use ADetailer, but there's still potential for further development ^^ that's why I prefer it for full body, especially if the style is more drawn ☺
Inpaint is the way to go
Try a higher base resolution as well. This model has no problems generating 1024x1536 images!
Congrats... It is an outstanding model... I love it.
thank u very much 4 ur kind comment ♥ I am very glad if u like Illustrij Quill ☺
That's a strangely precise hires multiplier (x 1,526).
That came out when I calculated what the hires settings were on site^^ so that I could show a similar quality for the showcase ☺
It was important to me to show directly how it can look like on site ^^
Of course, this is just an example of how I currently use it, which of course remains a matter of taste
Why the model unavailable?
hey, currently only a total of 100 models can be used on the site for on site generation and the prices per model are currently just over 30,000 buzz per week per model. Unfortunately I can't pay that for every model :I
I hope we have more options here someday too
@reijlita dang this one of my favs
love this model!
hi. i'm using your model as a refiner to turn family photos into 3d/anime-style images. using denoising strength about 0.15-0.3. i absolutely love it. thanks a lot for this great model.
Outstanding model.. even without lora, it makes extremely detailed and perfectly lovable pictures.