| --- |
| license: other |
| license_name: fair-noncommercial-research-license-v1 |
| license_link: https://huggingface.co/datasets/Rice-RobotPI-Lab/egoinfinity/blob/main/LICENSE-Action100M |
| pretty_name: EgoInfinity (preview) |
| task_categories: |
| - robotics |
| tags: |
| - egocentric |
| - imitation-learning |
| - retargeting |
| - manipulation |
| - action100m |
| size_categories: |
| - 1K<n<10K |
| --- |
| |
| # EgoInfinity (preview) |
|
|
| Derivative scene assets for a curated subset of Action100M (Meta FAIR) clips. |
|
|
| **Preview — not for general release.** Schema and contents may change. |
|
|
| ## License |
|
|
| FAIR Noncommercial Research License v1 (see [LICENSE-Action100M](LICENSE-Action100M)). Noncommercial research only. |
|
|
| Built by the **Rice RobotPI Lab**. |
|
|
| ## Retarget |
|
|
| For 104 of the clips under `samples/`, we provide retargeting results on four |
| robot embodiments under `samples/<clip>/retarget/<robot>/`: |
|
|
| - `franka` — Franka Panda (7-DoF arm + 1-DoF gripper) |
| - `g1` — Unitree G1 (7-DoF arm + 7-DoF hand) |
| - `robonaut2` — NASA Robonaut 2 (7-DoF arm + 18-DoF hand) |
| - `xlerobot` — XLeRobot (5-DoF arm + 1-DoF gripper) |
|
|
| Each `<robot>/` folder contains: |
|
|
| | File | Description | |
| |-------------------|---------------------------------------------------| |
| | `input_viz.mp4` | Ego input visualization used by the retargeter | |
| | `robot_sim.mp4` | Sim replay of the retargeted trajectory | |
| | `trajectory.npz` | Per-frame `q_left`, `q_right`, finger joints, joint names, fps, clip id | |
| | `torso_frames.npz`| Per-frame torso pose (`R_per_frame`, `t_per_frame`) plus anchor frame | |
| | `metrics.npz` | IK success rate, position/orientation error (per-frame, mean, converged-only), joint-limit margin, manipulability, roughness | |
|
|
| Per-clip frame counts range from 29 to 734; per-clip fps is recorded in |
| `trajectory.npz` (`fps` key) and is not uniform across clips. |
|
|
