Update!:Anima version is released. Trained on Japanese traditional art. Shunga etc.I made this lora with intention to generate anime-like images. Old Japanese art meeting modern art.Trigger word has become: "jptradartan artstyle".<hr />OLD (Illustrious):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.<hr />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 anatomy. 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 (huge breasts, 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.

Japanese Ukiyo-e / Shunga Style

Files with SHA256: a3e9d59318d8e23b3919879a036c43e96764cd575146a2659e63340c42f6b01c