This is a fork of H-Liu1997/BEAT2
that adds annotations contributed by the
RAG-Gesture (CVPR 2025)
and MIBURI (CVPR 2026) projects.
The base BEAT2-English data (motion, audio, TextGrids, semantic labels,
pretrained motion-autoencoder weights) is inherited verbatim from upstream;
the additional annotations from RAG-Gesture and MIBURI are pushed on top.
If you use only the original BEAT2 dataset, please cite… See the full description on the dataset page:
https://huggingface.co/datasets/m-hamza-mughal/beat2-additional-annotations.
