HumEnv
HumEnv benchmark
Simulated SMPL humanoid environment with reward, goal-reaching and motion-tracking test suites for whole-body control policies.1
Last checked 10 Oct 2026Basic entry11 of 15 facts checked at the sourceNext check 8 Apr 2027
Runs in
Simulation2
Checked against real robots
Not checked
Skill
Walking and balance
Robot
Game character1
Licence
Unclear3
Comparisons with real robots
Not checked4
Details
Simulated SMPL character only.
Details
About
- What it is
- Benchmark Inferred2
More
Classified by the Atlas from how the authors describe and distribute it.
- Built by
- Meta FAIR (same authors as Meta Motivo paper).1
- Released
- 2024-12-16 (tag v0.1.0).5
- Version
- v0.2.0.5
- Last update
- 2025-04-22, v0.2.0: fixes to benchmark distance-matrix metric and actuator range. Last push 2025-04-22.5
- Status
- No commits since 2025-04-22 (about 18 months before check date). Inferred1
More
Dormant by activity.
Setup
- Runs in
- Simulation Inferred2
More
Classified by the researcher from the primary page; no separate fact row.
- Robot
- Game character1
- Size
- Paper protocol: 45 reward functions, 50 manually selected goal poses, tracking on AMASS test split (990 motions, ~3 h; train 8,902 motions, ~29 h). Repo: 9 configurable reward classes.4
Scoring and access
Sources 6
- 1facebookresearch/humenv on GitHub (repository)Repository · checked 10 Oct 2026
- 2Zero-Shot Whole-Body Humanoid Control via Behavioral Foundation ModelsPaper · Apr 2025 · checked 10 Oct 2026
- 3facebookresearch/humenv on GitHub (blob)Repository · checked 10 Oct 2026
- 4Zero-Shot Whole-Body Humanoid Control via Behavioral Foundation Models (full text)Paper · Apr 2025 · checked 10 Oct 2026
- 5facebookresearch/humenv on GitHub (releases)Repository · checked 10 Oct 2026
- 6facebookresearch/humenv on GitHub (file README.md)Repository · checked 10 Oct 2026
Change history
- Created as a basic entry: identity facts checked at primary sources (phase 1 re-verification).