This lora was trained on about 50 images from the games series of Dragon Crown, Odin Sphere and a few other Vanillaware games. The goal of this lora was to capture the style. It's not prefect but it came out pretty damn close. I am satisfied.
V3: Trained on Noob 1.0 V-Pred.
V2: Trained on Illustrious and Zootlustrious.
To get the results you see in the images above, these are the settings I used.
Description
Trained on 50 images. The base for this lora is Noob 1.0 V-Pred.
This lora uses slightly different and improved settings to accommodate for V-Prediction. V-Pred improves prompt comprehension and color range greatly. Noob has a nice range of concepts compare to Illustrious 0.1. You need to make sure you're using the proper UI and settings in order to utilize V-Pred. I don't expect this to work on Eps. models like Illustrious 0.1.
FAQ
Comments (4)
Awesome style!
Given the softness of the style, I suspect there might've been a good dose of Unicorn Overlord in there.It's not really a criticism, just a matter of taste, but damn I do miss the sterner shading and lines from George Kamitani's art in DC and Muramasa Rebirth.
There's actually zero Unicorn Overlord images used. All the data is from Dragon Crown and past illustrations before then. The style of all of the images aren't very uniform, so the lora is a unconventional blend of Vanillware. I imagine if I just focused on the Dragon Crown images I could get a more accurate recreation.
@Fish788 Oh wow, there I go making assumptions then.