This is a lora for Hunyuan Video generative text-to-video model.
This was trained on various videos of people performing the same walking animation.

It is generalized to work with any character/subject

trigger: "kxsr"
suggested prompting style: "kxsr, {your character name}, full body, no_crop"

use at strength of 1.0

If doing animated characters, do strength of 0.85-1.0
If doing realism celebrities, do strength of 1.10

The Hunyuan Video Wrapper by Kijai now natively supports the lora loader and you should use it.