Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

Motius Leaderboard Case Assets

This dataset stores compact browser assets for the all-case comparison pages in the public Motius leaderboards. It is a visualization companion, not a training or evaluation dataset.

Folder Population Comparison
m2t-humanml3d-smpl/ 4,400 HumanML3D input motion with TM2T, MotionGPT, MotionGPT3, and VerMo captions
t2m-humanml3d-smpl/ 4,042 HumanML3D selected captions with every released T2M output
babel-sequential-smpl/ 1,295 BABEL GT, FlowMDM, MotionStreamer, MotionLab, and PRISM episodes
temporal-condition-smpl/ 4,012 per setting Six temporal-control settings with GT and five generation methods

Motion pages decode the repository-standard SMPL motion_135 representation: root translation plus 22 row-major rotation-6D joints. Translation is quantized per clip to unsigned 16-bit values, while rotation-6D channels use a fixed signed 16-bit scale. The Three.js viewer applies these transforms to a neutral SMPL template with linear blend skinning, so the displayed result is an animated body mesh rather than a joint-only skeleton.

Each manifest records source frame count, playback rate, stride, byte offsets, and translation quantization parameters. All methods for a selected case are loaded together and played on one synchronized timeline.

Assets are generated by:

python tools/export_smpl_web_model.py --help
python tools/build_smpl_motion_gallery.py --help
python tools/build_m2t_smpl_gallery.py --help
python tools/build_temporal_case_explorer.py --help
python tools/build_babel_smpl_gallery.py --help

Metrics never consume these quantized browser assets. Evaluation uses the original full-precision motions; this dataset is only for qualitative review.

Downloads last month
-