:3
I'm rewatching the donghua and refining her training data. It might take some time, but I'll create another version.
When I made the LORA, I wasn't using RTX video enhancement, so many image qualities were poor. It also didn't help that Zhu Zhuqing rarely appears full-body, and when she does, it's for a second, leading to many blurry images that lost characteristics during upscaling.