Oborune - a style blend led by ukiyo-e, with anime illustration on aged paper and nihonga.
Trigger: oborune,
oborune, fine linework, flat colors, soft shading, intricate details, balanced composition,Base model: Illustrious
Purpose
The goal of this LoRA is not to reproduce ukiyo-e or nihonga as such. It carries over what makes them recognizable - the palette, the contrast and the painterly feel - while keeping the image an anime illustration.
Weights:
1.0 - general use, balanced blend.
1.3-1.5 - for results closer to ukiyo-e or nihonga. Note that hands and faces may develop artifacts at these weights; ADetailer compensates for this.
Settings
Sampler: Euler a.
Steps: 30-40
CFG: 4.5-8
Negative: signature, japanese text, border, watermarkFor faces and eyes I recommend the ADetailer extension - https://github.com/Bing-su/adetailer
Happy generating, and may your seeds be lucky! :)
If you enjoyed this work, please leave a like - it really keeps me motivated :D
Description
oborune, fine linework, flat colors, soft shading, intricate details, balanced composition,
Comments (3)
I am thoroughly enjoying this. Thanks for creating. Thoughts for creators: the lora appears weighted toward Japanese clothing, poppies, and chrysanthemum as defaults. If you don't specify, this is often what you'll get.
Thank you so much for such a thoughtful review!
Yes, the issue with the clothing comes from the dataset itself. Unfortunately, I couldn’t find enough suitable images with more varied outfits without losing the original concept and style I wanted to achieve.
Still, thank you for your helpful and kind feedback :)
Interesting style. It seems to work best with the checkpoint used by OP ( IllustriousNXT_XL). Results may vary with other checkpoints.