A small Tendou_Aris character enhancing LoRA for MOSS Image. Given that the MOSS Image model is already quite powerful on its own, this LoRA focuses primarily on preserving specific character traits—such as face shape and the halo—during generation; therefore, you should not expect it to produce drastic changes. Additionally, there are still some limitations regarding the rendering of the irises, as it fails to accurately replicate the white circular rings characteristic of Aris’s eyes.
Training data based on the SFW sources, chant and WONT make sure it'll work well on NSFW content

