We had an outage July 5th where users were not compensated, a fix has gone out.

July 5th Compensation

Krea 2 LoRA training is 50% off, and we're running a Krea 2 Training Contest with 1.8 million Buzz in prizes across four LoRA categories.

⚡ Krea 2 Training Contest + 50% Off Training

I noticed a general lack of general anime cel shading style models (outside of v-tuber specific ones) so I decided to train one. It was trained on images generated using 3D modeling software with a cel shader on a variety of subjects, but primarily humans. It seems to be pretty general in what it can generate while still maintaining the style. Check each version's release notes for more info on what it is capable of and what issues it may have.

Flux version of the model trained at 1024 res on civitai's training service on the same data set as prior versions (30 images). Seems to work equally well with sentence prompts and tag prompts. The only tag you need to trigger the style is "cel_shading".

FLUX

SDXL version trained on the same dataset as the pony version using civitai's training service. This version is slightly less anime-esque due to lower training epochs, however it still consistently produces a cel shading style on the variety of SDXL models I have tested it on. I recommend a strength of around 1.0. The trigger word is simply "cel shading", and if you want a more consistent anime style, you can also use "anime".

SDXL 3 Epochs

Pony XL version of the model trained using Civitai's training service for 7 epochs on 30 images. I tested it on a variety of pony checkpoints and it seems to produce results best on more 2d checkpoints, but you can get it to work on photorealistic pony checkpoints with the right amount of negative keywords. Seems to work best at a strength of 1, you can tweak it if you have difficulty getting the style to work properly or think its influencing your subject too much. The trigger for the style is simply "cel shading".

PonyXL

Stable Diffusion 1.5 version trained on around 30 images of various subjects on Kohaya SS at 512 resolution. The trigger for the style is just "cel shading", and I trained it on booru style captions, so those should work best to describe the subject. It seems to work in a strenght of around 0.7 to 1.0, with 0.85 seeming to produce the style consistently while anything over 0.9 might start to impact the subject.

Stable Diffusion 1.5

Cel Shading Style (CGI Anime)

physical violence

weapon violence

wide hips

revealing clothes

downblouse

convenient censoring

pg-13

corpses

suggestive

oral invitation

pg13

sexy

huge breasts

thick thighs

sexual situations

male nudity

disturbing

male swimwear or underwear

female swimwear or underwear

partial nudity

undressed

female nudity

breasts out

exposed female nipple

breast out

lingerie

male underwear

hair over breasts

female swimwear

gigantic breasts

no panties

graphic violence or gore

covered nipples

huge butt

strapless leotard

sitting on face

emaciated bodies

one breast out

female underwear

nude

nsfw

graphic male nudity

adult toys

illustrated explicit nudity

nudity

graphic female nudity

hentai

futanari

porn

sexual intent

genitals

peeing

vore

oral

sexual activity

anal

blowjob

dildo riding

incest

hanging

hate symbols

nazi party

white supremacy

diapers

scat

self injury

hate speech

urine

extremist

child on child

latex clothing

swimwear

bukkake

fellatio

cumshot

implied fellatio

eat_cum

cumdrip

cum in pussy

cum on face

after fellatio

cum on hair

cum on body

cum on tongue

cum on hands

cum in mouth

triple fellatio

autofellatio

fucked silly

cum on pussy