Sign In

DarkfireStyle35 - YMMV Diceroll, Read the Description

184

Updated: Feb 5, 2026

style

Not loaded

Download

1 variant available

SafeTensor

162.67 MB

Verified:

Type
LoRA
Stats

73

184

133

Generation

Not loaded

Reviews
Published

Feb 5, 2026

Base Model

Illustrious

Training
Steps: 2,000
Usage Tips
Clip Skip: 2
Hash
AutoV2
196BCB178D
default creator card background decoration
Followers - 2185

2.2K

Likes - 2732

2.7K

00017-1924095509.png

This model was an experiment. Instead of a carefully curated set of aligned images it was built from 150 high quality unrelated illustrations. Mostly fantasy and sci-fi, but anything goes. Only rule was I liked it and it had good detail. Generations are 50-70% like the posted showcase, but can jump all over the place.

Additionally I think this model needs hires fix applied. Every image I generate runs hires fix in Forge. There is an option to do this with on site generations. It's been a while since I tested, but I believe it's an option after images are created, costs a bit more buzz for the second round of processing, and by default upscales by 1.5x which is the same I do.

Since this model was random randomness and not a curated style of images I put together it does not code for skin tone the way my other recent models have.

All that said, if you make something nice with it, please share!