Sign In

AestheticBunny Artstyles

92

1.1k

193

46

Verified:

SafeTensor

Type

LoRA

Stats

62

2

Reviews

Published

Jul 23, 2025

Base Model

Illustrious

Training

Steps: 2,146
Epochs: 33

Usage Tips

Clip Skip: 2
Strength: 1

Trigger Words

retpixcol
colorful
abstract
multicolored background

Hash

AutoV2
6863D19D35

Going to put various new styles here overtime, my first one being an illustration style. These are just made with art by various different people. Clip Skip 2

Retro Pixel Colour
Lora Weight 1 (This style is painful to look at at times but has a bit of charm)
Trigger Word: retpixcol
Optional Tags To Try: colorful, abstract, multicolored background
(Note: 'colorful' and 'multicolored background' are what I think really trigger the bright colors. If you don't want this, don't use them.)

Comic V1-A
Lora Weight 1 (or 1.2 which is maybe better stylistically but also a bit more incoherent.)
Trigger Word: comicaesbun
Does a superhero comic sort of style kinda.

Dramanime
Lora Weight 0.9
Trigger Word: dramanime
This lora focuses on anime with dramatic lighting, for some things you might be good increasing the weight to 1.2 perhaps. The effect is very visible at 1.4, however it also seems to make things a little messy at that point... Relatively smaller dataset I used and there are a fair bit of watermarks in there so I'd suggest using the negative below. Had a good bit of fun with it so there'll be a good amount of images I generated which I'll share.
Negative: watermark

Illustration B-V1
Weight is tricky, I'd say between 1.4 and 1.8 with 1.8 being heavy (possibly too heavy? Cool if it works though), 1.6 being moderate and a bit more reliable, and 1.4 being a lighter effect that is the most reliable. I think I like 1.8 when it works but if you want a general recommendation maybe more 1.6 for more consistency in quality.
Trigger Word: illustrationBV1
(Note: Very different sort of illustrations from illustration A-V1. Completely different dataset.)

WarcraftCinema
Weight 1.4 (at least for the model I used, might vary though).
Trigger Word: warcraftcinema
Possible Negative: long eyebrows
(Note: Trained on cinematic sort of edits of World of Warcraft models. Sometimes gives blood elf brows to ordinary characters, but having 'long eyebrows' in the negatives can help with this. I also trained it to include different Warcraft races, using tags like 'undeadmal' or 'humanmal' for male undead and humans, and 'voidelffem', nightelffem' and 'succubusfem' for females of those races. Others too, I didn't test them all since the focus was on the style. Still, if you want a particular race it might be better to just describe them with your prompt.)

Neoventure Toon
Weight 1.2 gets the style a bit more, lower to 1 potentially.
Trigger Word: neoventuretoon
(Note: This one's a bit focused on capturing a sort of Neopets/Adventure Quest sort of vibe with the art. There isn't too much of that in the dataset but that's what I was going for.)

PixelLite
Weight 1 is good, probably no higher than 1.6. Increasing it from 1 works better with hi-res fix.
Trigger Word: pixellite
(Note: I also recommend 'pixel art' as a tag. This is my first pixel style lora and I might do more in the future in different styles but this is just focused on a light pixelated sort of style, nothing to heavy or anything hence the name. You can make it more or less heavy though somewhat via adjusting the weights. For some concepts 1.6 might work better than 1. For reference the images I shared that start off with the blonde elf were done without hi-res fix and has a more significant pixilation.)

MedievalWave
Very adjustable. Weight of 1 works but can raise or lower it to suit needs.
Trigger Word: medievalwave
(Note: This style was kinda focused on an aesthetic with long wavy hair that has strokes between it. The end result though is a pretty interesting style. Seems to add a lot of general intricate details to an image I find.)


Sapphrost
1.3 weight gives about what I was aiming for.
Trigger Word: sapphstyle
Variant 1: sapphfrost
Variant 2: sapphdepths
Variant 3: sapphsea
Variant 4: sapphforest
Variant 5: sapphfocus
(Note: Bit different. Pick a variant and add it after the trigger word to mix things up a little. Also recommend the "blue theme" tag for all variants. "snow" can be a good tag for the saphhfrost variant. "underwater" can be good for sapphdepths. Sapphsea is more aimed around things entailing "ocean" especially stormy oceans. Sapphforest could work well with trees or maybe mushrooms. Sapphfocus is just another option for a neat bluish character.)

Gothic A-V1
0.75-1.2 weight are really good.
Trigger Word: GothicAesBun
(Note: Recommend "colored skin, grey skin" as it fits lots of the training data. Also does a lot of pointed sort of chins. Might give multiple panel images, I'd suggest in the negatives "borderless panels, multiple views, segmented comic" to fix. Also if backgrounds are simple, add "simple background" to negatives or just describe a background and that should work.)

Illustration A-V1
0.75 seems like a good weight to me.
Trigger Word: IllustrationAesBun
(Note: Seems to like doing upper bodies somewhat and sometimes different art panels for some reason. Seems like it might get a bit more 2.5D with stronger weight?)