The training image is the same as V3.0, but more animated and contrasty images are more easily generated.
This LoRA is used as a base for learning images when creating LoRA and OpenPose.
