Type | |
Stats | 570 1,690 |
Reviews | (69) |
Published | Nov 24, 2023 |
Base Model | |
Usage Tips | Clip Skip: 2 Strength: 0.8 |
Hash | AutoV2 D9BF730F4A |
A Lora trained with (at the moment) 50 images, will add more if/when i update it.
Helps with non-contemporary images.
got the idea after seeing some Malazan fanart (good book series btw, atleast imo)
works best at fantasy images but has no issue doing contemporary/futuristic. works with 2D and 2.5D models, not sure about 3D. Biggest issue so far is that in landscape ratios it likes to gen multiple people.
weight 0.7 - 1.0 ( 0.7 - 0.8 is probably fine, going higher is possible, but may distort characters).
I recommend using "drawn in the artstyle of <Lora>" somewhere in your prompt ( i use it at the end) it's not necessary, but it does change the image quite a bit with/or without it.
Helpful tags: dark fantasy, depressing, dark, creepy, dark theme,
Optional solo, solo focus, (These may help along with the negatives if it gens multiple people)
Helpful Negative: silhouette, multiple people