This model was trained from Civitai, will keep update the model in future.
Version 2.1 has been released.
Changes made in version 2.1:
refine in prompts for more accurate results
add training image samples
Better result for weapon: rebound ball (trigger prompt: holding rebound ball)
I'm still figuring out how to combine with ep19 model, the combines model will be released soon.
Same as previous versions, please use facefix for better quality / results. (Recommended Strength: 0.8)
I'm wondering how to reveal the details on the skirt part, sometimes it generates full white skirt, sometimes it generates full red skirt. (But most of the case can reveal the details so this is minor problem)
Feel free to leave your comments, your thoughts on the model.
Description
Update Notes on v2:
Trained by Pony Diffusion v6
155 Upscaled screencaps used (Mostly focusing on Supervising Animator A, others are supporting character details)
Redesigned on trigger words, support on using weapon (rebound ball)
(I guess) Can revealed about 90% of the character's features (including side and back)
Face fix is strongly recommended for this LORA.
Other trigger words you may like to use:
Demon: Purple Aura, energy attack
Rebound ball: Rebound ball, holding rebound ball / throwing rebound ball
This model was built through a lot of trial and error. I would like to thank everyone who downloaded my previous work and the guidance from doujins, who encouraged me to keep working on LORA development. I really love this character very much and I finally make this LORA.
Will keep update on the LORA model. Feel free to comment on the LORA model.