Trained on AnyLora
To get the best results, please read the description below:
Recommended weight: 0.7-0.9
Activation word: Viola
Use of " beautiful detailed eyes, " and "Detail Tweaker LoRA" models is also recommended for good results
Upscale to 1.5
Using anime SD models is recommended, i haven't tested semi-realistic ones
I would appreciate any reviews, positive or negative!
I would very much appreciate you sharing your results ^_^
If you have any suggestions on what LoRA i should do next, please share them in the comments
Enjoy :)
Description
Version 1 - trained on 51 images
FAQ
Comments (3)
Interesting - all of your preview images (for all of the OP loras) are pretty heavily stylized. Do these style attributes leak into the character loras, or are they clean and the images are just based on unrelated model output / settings?
This particular style appears mainly when i use Anything v4.5 with tags such as (cgi, film grain, RAW Photo, e.t.c.). Removing this from the prompt and selecting a different Checkpoint such as AnyLora(on which all anime models are trained) would change this style to a more generic one.
I present images in different styles in few of my models (See my Nami model, for example), but i don't bother doing it on all of them because it takes a bit too much time tbh :)
I use this style for preview just because i personally like it.
So no, this style does not leak heavily into character LoRA's as far as i have tested them!
Sounds good, thanks for the clarifications!