








Update!:
NoobAI (illustrious) version is released. I used a larger dataset this time.
From day / night cycle to rain, seasons ( fall - winter) and many others.
I solved all the problems Pony and SDXL versions had.
Actually, this lora is so powerful.
VERY IMPORTANT So i recommend using it at 0.3 ~ 0.5 weight.
If the character lora you are using is too powerful, you can go higher,
(Raiden Shogun lora is mine. And it's too powerful for example.)
but these weights are the best in most cases.
OLD [Pony / SDXL]:
Trained on Japanese traditional art. Shunga etc.
I made this lora with intention to generate anime-like images.
Usage without the trigger word could result in more anime-like output.
The case i will explain below can be by-passed with increasing the weights of the lora like "jptradartpny:1.5" in expense of ???????. Meaning by increasing the weights too much, deformed body parts will be more common. Anyway, there is a possibility of:
If you type exaggerated bodies (????????????, thick thighs etc.) The result could be more anime-like.
If you type any hair color other than black, any eye color other than black-brown, the result could be more anime-like.
1. 轉载模型僅供學習與交流分享,其版權及最終解释權归原作者。
2. 模型原作者如需認領模型,請通過官方渠道联系海藝AI工作人員進行認證。我們致力於保護每一位創作者的權益。 點擊去認領
