sexual situations

physical violence

disturbing

male nudity

extremist

hanging

hate symbols

nazi party

revealing clothes

weapon violence

female swimwear or underwear

male swimwear or underwear

graphic female nudity

corpses

peeing

emaciated bodies

female nudity

undressed

male underwear

female swimwear

genitals

female underwear

thick thighs

breasts out

strapless leotard

vore

bestiality

breast out

one breast out

huge breasts

gigantic breasts

huge butt

covered nipples

hair over breasts

no panties

sitting on face

anal

dildo riding

downblouse

oral

porn

futanari

hentai

nude

lingerie

sexy

nsfw

partial nudity

white supremacy

adult toys

graphic male nudity

illustrated explicit nudity

nudity

self injury

graphic violence or gore

wide hips

convenient censoring

oral invitation

exposed female nipple

blowjob

sexual activity

sexual intent

suggestive

bukkake

cumshot

cumdrip

cum in pussy

cum on face

cum on hair

cum on body

cum on tongue

cum on hands

cum in mouth

cum on pussy

eat_cum

fellatio

implied fellatio

after fellatio

triple fellatio

autofellatio

pov fellatio

fucked silly

presenting ass

presenting pussy

spreading legs

spread legs

swimwear

Please go an appreciate the original artist DOKA <a rel="ugc" href="https://www.pixiv.net/en/users/36455908">ドカ</a>promts are: ama rei, by doka, ... (not sure if you can leave them, didn't test)Weight 1.0 seems to be working best. Feel free to test out different weightsevery second picture in my example is without lora for comparison :)

New Version applys the style way better with a hell of lot more details.I changed the way i will train my loras now. it seems using repeats to increase the steps for training works way better than having high epoch numbers.Dataset is still 41 images. Settings i used now are: Batch 4, repeats 5, steps 2450 (9800 without batch), epoch 50, lora rank 32

v1.0

version 0.5 i did localy on my computer (RTX4080 super, ryzen 7 7800X3D) using the program OneTrainer using the guide from <a target="_blank" rel="ugc" href="https://civitai.com/articles/4789/lazy-lora-making-with-onetrainer-and-ai-generation">Phil866</a>. 100 Epoch by default and batch number 4 Dataset of 41 images. seemes to be a little bit to less.i might do another version with increased steps/epochs since its not applying the style as hard as expected.

v0.5

Doka ドカ (Artist Style) [PonyXL]

Type	LoRA
Stats	541 76
Reviews	Very Positive (74)
Published	Apr 10, 2024
Base Model	Pony
Training	Steps: 2,450 Epochs: 50
Usage Tips	Clip Skip: 2 Strength: 1
Trigger Words	ama rei, by doka,
Hash	AutoV2 A5F70A7705