physical violence

revealing clothes

weapon violence

pg-13

corpses

wide hips

convenient censoring

oral invitation

thick thighs

huge breasts

downblouse

suggestive

sexy

pg13

sexual situations

disturbing

male nudity

female swimwear or underwear

male swimwear or underwear

partial nudity

graphic violence or gore

emaciated bodies

exposed female nipple

female nudity

undressed

male underwear

female swimwear

female underwear

breasts out

strapless leotard

breast out

one breast out

gigantic breasts

huge butt

covered nipples

hair over breasts

no panties

sitting on face

nude

lingerie

nsfw

adult toys

nudity

graphic male nudity

illustrated explicit nudity

graphic female nudity

sexual intent

genitals

porn

futanari

hentai

peeing

blowjob

sexual activity

vore

anal

dildo riding

oral

incest

hanging

hate symbols

nazi party

white supremacy

self injury

extremist

hate speech

diapers

urine

scat

child on child

bukkake

fellatio

bikini

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

A versatile model for creating icon art for computer games that works in multiple genres and at various levels of detail.I'm happy with where this is, so there won't be further updates until there's a new major version of Stable Diffusion.

VERY IMPORTANT: FOR THIS TO WORK WITH AUTOMATIC1111, BE SURE TO ALSO DOWNLOAD THE CONFIG FILE AND PLACE IT IN YOUR MODELS FOLDER ALONGSIDE THE MODEL. This may change in the future as automatic1111 support for 2.x matures.This is a finetune of Stable Diffusion 2.1 (as opposed to 1.4), and the results it produces are way more interesting. Click the 'i' icon on the example images of how to prompt it (negative prompt is important here). If you feel like it, please experiment and leave any interesting discoveries in the reviews.License: I don't believe models ought to be copyrightable, so you can do whatever you want with this, including uploading and redistributing it elsewhere, finetuning it, berrymixing it, etc. I do humbly request that you share any models that you make with this one, but you're free not to.This model was finetuned with <a href="https://github.com/devilismyfriend/StableTuner" rel="ugc" target="_blank">StableTuner</a> with a dataset of ~250 manually captioned images. Learning rate was 1e-6, 250 epochs, batch size of 24. Text learning cutoff was 75 epochs to prevent overfitting.Note: I've accidentally discovered that this model does some amazing non-icon art. See the cityscape picture for an example of another way to prompt it.

v4 SD 2.1

v2 was an overfitted mess, so I'm not bothering to post it here. :)This version produces more consistent, varied, higher quality output, and is a lot better at abstract ability icons. It also requires shorter prompts to produce good output. Click the 'i' icons on each image for prompt and settings.I find that experimenting with different artist names doesn't replicate those artists' style, but it can add more detail.For the next version of this, I may try finetuning Stable Diffusion 2.1 512x512 instead, to see if that helps it a bit with more complicated prompt comprehension.

It's pretty good at objects, but has trouble with abilities. I'm going to try and retrain it with some more ability icons added to the set. Works best if you append ", stylized game icon" to your prompt. See "i" icon on individual images for details. Examples are cherry picked from a larger set.Trained with <a href="https://github.com/devilismyfriend/StableTuner" rel="ugc" target="_blank">StableTuner</a>. This model is based on SD 1.4.

Stylized RPG game icons

Type	Checkpoint Trained
Stats	569 15 12
Reviews	Positive (14)
Published	Dec 9, 2022
Base Model	SD 1.5
Training	Steps: 24 Epochs: 100
Trigger Words	stylized game icon sci-fi steampunk spell
Hash	AutoV2 B5E6567AC6