OmniHuMo

Overview

👁 OmniHuMo Dataset

OmniHuMo comprises over 3.2 million high-quality motion-capture sequences — sourced entirely from internet videos — totaling more than 5,000 hours.

Get Started

# Ensure git-lfs is installed (https://git-lfs.com)
git lfs install
# When prompted for a password, use an access token with write permissions.
# Generate one in your settings: https://huggingface.co/settings/tokens
git clone https://huggingface.co/datasets/L-yiheng/OmniHuMo
# To clone without large files (only their pointers)
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/datasets/L-yiheng/OmniHuMo

Dataset Processing

Folder Hierarchy

.
|-- README.md
|-- assets
|-- audio.tar.gz.part_aa
|-- audio.tar.gz.part_ab
|-- audio.tar.gz.part_ac
|-- audio.tar.gz.part_ad
|-- audio.tar.gz.part_ae
|-- audio.tar.gz.part_af
|-- audio.tar.gz.part_ag
|-- audio.tar.gz.part_ah
|-- audio.tar.gz.part_ai
|-- audio.tar.gz.part_aj
|-- audio.tar.gz.part_ak
|-- audio.tar.gz.part_al
|-- audio.tar.gz.part_am
|-- audio.tar.gz.part_an
|-- audio.tar.gz.part_ao
|-- audio.tar.gz.part_ap
|-- audio.tar.gz.part_aq
|-- audio.tar.gz.part_ar
|-- audio.tar.gz.part_as
|-- audio.tar.gz.part_at
|-- audio.tar.gz.part_au
|-- audio.tar.gz.part_av
|-- audio.tar.gz.part_aw
|-- audio.tar.gz.part_ax
|-- audio.tar.gz.part_ay
|-- audio.tar.gz.part_az
|-- audio.tar.gz.part_ba
|-- audio.tar.gz.part_bb
|-- audio.tar.gz.part_bc
|-- audio.tar.gz.part_bd
|-- audio.tar.gz.part_be
|-- audio.tar.gz.part_bf
|-- audio.tar.gz.part_bg
|-- audio.tar.gz.part_bh
|-- audio_feat.tar.gz.part_aa
|-- audio_feat.tar.gz.part_ab
|-- audio_feat.tar.gz.part_ac
|-- audio_feat.tar.gz.part_ad
|-- audio_feat.tar.gz.part_ae
|-- audio_feat.tar.gz.part_af
|-- audio_feat.tar.gz.part_ag
|-- audio_feat.tar.gz.part_ah
|-- omnihumo_v0.tar.gz.part_aa
|-- omnihumo_v0.tar.gz.part_ab
|-- omnihumo_v0.tar.gz.part_ac
|-- omnihumo_v0.tar.gz.part_ad
|-- omnihumo_v0.tar.gz.part_ae
|-- omnihumo_v0.tar.gz.part_af
|-- process_code
|-- split.tar.gz
`-- upload.py

Decompression

cat ./audio_feat.tar.gz.part_* | pigz -d -p 64 | tar -xvf -
cat ./audio.tar.gz.part_* | pigz -d -p 64 | tar -xvf -
cat ./omnihumo_v0.tar.gz.part_* | pigz -d -p 64 | tar -xvf -
tar -xzf split.tar.gz

License and Citation

All the data and code within this repo are under CC BY-NC-SA 4.0. Please consider citing our project if it helps your research.

@misc{li2026anymoscalinganymodalityconditional,
 title={AnyMo: Scaling Any-Modality Conditional Motion Generation with Masked Modeling},
 author={Yiheng Li and Zhuo Li and Ruibing Hou and Yingjie Chen and Hong Chang and Hao Liu and Shiguang Shan},
 year={2026},
 eprint={2605.29488},
 archivePrefix={arXiv},
 primaryClass={cs.CV},
 url={https://arxiv.org/abs/2605.29488},
}