# nghorbani/amass

Data preparation and loader for AMASS

Repository: https://github.com/nghorbani/amass
Canonical: https://ross.abutalabs.com/products/nghorbani-amass
Homepage: https://amass.is.tue.mpg.de/
Language: Jupyter Notebook
License: NOASSERTION
License Family: other
Topics: human, motion, action-recognition, pose-estimation, motion-capture
Last push: 2024-07-25T11:10:05+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2581, "days_push": 769, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1010, forks 102 (observed 2026-08-28T04:03:13.093252+00:00)

## What it is
AMASS is a large unified database of human motion capture data representing 15 optical marker-based mocap datasets in a common SMPL+H body model parameterization. This repository provides data preparation tools, loaders, and Jupyter notebook tutorials for using AMASS in research, animation, and deep learning training.

## Use cases
- load human motion capture data for deep learning
- animate SMPL or SMPL-X body models from mocap data
- prepare train/validation/test splits of 3D human motion
- visualize 3D human body meshes from motion data
- generate synthetic mocap training data
- research human pose estimation and action recognition

## When to choose
- you need large-scale unified 3D human motion data for training motion models
- you work with SMPL/SMPL+H/SMPL-X body models in PyTorch
- you need consistent mocap parameterization across multiple source datasets

## When to avoid
- you need raw marker-based mocap data rather than SMPL-parameterized poses
- you need a non-Python or non-PyTorch pipeline
- your use case is outside research licensing terms

## Facets
- artifact type: dataset
- maturity: stable
- function: data-science, machine-learning, data-visualization, etl
- domain: computer-vision, machine-learning, graphics, simulation
- platform: python, cross-platform
- tags: motion-capture, human-pose, smpl, body-model, 3d-human-motion, animation, dataset-loader

## Member repositories
- nghorbani/amass (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:13.093252+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:11:57.708253+00:00, confidence not recorded.
  - readme: https://github.com/nghorbani/amass (fetched 2026-08-28T04:03:13.093252+00:00, sha 469bae7a6115)
  - homepage: https://amass.is.tue.mpg.de/ (fetched 2026-08-29T13:12:00.754418+00:00, sha 613904ec5793)
- Data as of 2026-08-30T08:39:29.467469+00:00.