V2-il

tagger：

kakenari_style, PULIN_nabe,

negative prompts：

uneven eyes,

Because the eyes original artist draws are basically asymmetrical, it may generate asymmetrical eyes. If you care, please add 'uneven eyes' to the negative prompt.

Increase more training databases, fix overfitting issues, labels, mosaics, dialogue bubbles, and other problems. Enhance art style.

add a little furry data.

Pony ver:

prompt: cel-shaded style,kekanari_style,

pony version is much bigger,but stabler,I train it with higher training parameters. and support you guys use 'homsimile pony ver' to txt2img, it's easy to use,.

if u add "greyscale',you can get comic book-style effect as cover

and suggest this embeddings:

https://civarchive.com/models/831971?modelVersionId=930629

Art style:

suggested strength:0.7-0.9

tag: kakenori karenari

It's art style, it'd better be mixed to other character lora, or Character prompt.

if you find the too high strength influenced your pic，or strength of other LoRA influenced the style, add ‘2010s’ or '2010s \(style\) ', you can control the art style much easier and more effectively.

no character lora

with character lora(strength 0.7):

with popular character prompt:

please support original artist: kakenari - pixiv

The model is for communication and learning purposes only.

Please notufy me to delete it if it infringes on any rights.