thick thighs

sexual situations

physical violence

disturbing

extremist

hanging

hate symbols

nazi party

breasts out

strapless leotard

revealing clothes

weapon violence

emaciated bodies

female swimwear or underwear

male swimwear or underwear

partial nudity

self injury

white supremacy

adult toys

graphic male nudity

illustrated explicit nudity

nudity

sexual activity

graphic violence or gore

graphic female nudity

vore

bestiality

corpses

breast out

one breast out

huge breasts

gigantic breasts

huge butt

wide hips

covered nipples

hair over breasts

no panties

sitting on face

convenient censoring

anal

dildo riding

downblouse

hentai

porn

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

lingerie

bikini

latex clothing

TRPCError

Liselotte Cretia from Seirei Gensouki trained on mostly official light novel art.Trained it from scratch using new settings and manually processed some images in the dataset. Especially hair should be much more consistent now compared to previous versions.

Fixed hair bow captions in dataset (white hair bow -&gt; white bow, hair bow) as NAI with clip skip 2 doesn't seem to understand that white hair bow is white bow rather than white hair, it was spoiling previous version. Actually noticed it before previous release with red bow but forgot to change captions with white bow as well.Also, changed LR scheduler in training settings from constant to cosine. (with prodigy optimizer)It's more accurate with 7 epochs compared to 8 epochs in previous version.

v2.1

v2.0

LoCon trained with the same image set and number of steps as LoRA. Superior with both quality and fidelity from what I can see in generated images. If hands are bad, easynagative embedding with value &gt;1 can help a lot without decreasing quality. I recommend 0.6-0.7 weight.

LiselotteCretia_LoCon_48epochs

Still behaves well after twice as much epochs. Far inferior to LoCon model, use that one instead if you can. (Going to upload it in a few hours after this one)

LiselotteCretia_LoRA_48epochs

Previous LoRA trained 12 epochs further with the same settings, still behaves well with high weigh

LiselotteCretia_LoRA_24epochs

This version is trained without pruning captions, so it needs a lot of trigger words to create Liselotte consistently. It tends to work less consistent without <code>hair bow</code> in particular. Trained with NAI (12 epochs x 18 images x 15 steps), examples generated with AOM3a1. All examples used openpose contronets and upscaled with multidiffusion plugin.

LiselotteCretia_LoRA_12epochs

Liselotte Cretia (Seirei Gensouki)

00148-1823150470-pfp.png

Type	LoRA
Stats	398 70
Reviews	Very Positive (62)
Published	Mar 22, 2023
Base Model	SD 1.5
Training	Steps: 270 Epochs: 24
Trigger Words	liselottecretia long hair wavy hair bangs blue eyes blue hair frills capelet pleated skirt shirt +1 more
Hash	AutoV2 A5EEA3042D