Model from Huggingface doesn't support wholebody #165
Description
Thank you for your excellent work. When I try to infer videos by VitPose+ Huge/Large models based on Huggingface, even if I set the dataset_index=5, the output keypoints are still 17. Do you know why?