Sign In

Soggy Skeletons (XL Style/Char LORA)

44
310
7
Verified:
SafeTensor
Type
LoRA
Stats
310
0
Reviews
Published
Jul 31, 2023
Base Model
SDXL 1.0
Training
Steps: 1,500
Epochs: 6
Usage Tips
Clip Skip: 2
Trigger Words
ssk skel
Hash
AutoV2
CE873FF041
default creator card background decoration
ediboo's Avatar
ediboo

First time training anything, started creating some random images on an app and really liked where things went so thought I would train a model based on the images. It basically generates images of skeletal figures in a rainy/wet setting in a cartoony or painted style. This is a very odd and specific theme so I'm not sure many people will like it. It was only trained on images of feminine/female figures, I used a tutorial to train it and have not tested it 100% thoroughly.

Sample prompt:

portrait shot of (ssk skel), hdhq <lora:SskSkel-v1:0.3>

Sample settings:

sd_xl_base_1.0 checkpoint, DPM++ 2M Karras or Euler A, 25-30 steps, cfg 7-9, clipskip 2-4, 1024x1024 or 832x1216 or 1216x832

I'm not sure if hi res fix or i2i refiner is needed as it seems to produce what I was going for but feel free to experiment and use it however