Mobjaverse

Mobjaverse: A Large-Scale Rigged 3D Model Dataset with Skeletal Animations Mobjaverse is a curated dataset derived from Objaverse-XL, specifically designed for research on skeletal animation understanding, motion generation, and articulated 3D shape analysis. It is curated in the paper TopoCap: Learning Topology-Agnostic Motion Priors for Monocular Video-to-Animation. Mobjaverse contains ~19k rigged 3D models spanning ~5k distinct skeletal topologies and ~2M motion frames… See the full description on the dataset page:

Source: Hugging Face Hub (duckduckplz/Mobjaverse). Metadata imported from the dataset’s Hub tags.