Source: https://actionable.ai/datasets/R01.html

[Skip to content](https://actionable.ai/datasets/R01.html#main)

ROBOT / MULTI-VIEW / AA R01

# Real Robot · G1 / G2

Whole-body and dual-arm experience from robots in real environments.

[Enquire about this dataset ↗](https://actionable.ai/?dataset=R01#contact)[Browse all datasets ↗](https://actionable.ai/#catalogue)

COLLECTION INVENTORY20,000 hours

SAMPLE RECORDINGS53

Sample access, licensing, and delivery are confirmed for each enquiry.

CAPTURE DETAILS

## Sample specifications

| Capture system | AGIBOT Genie-01 / Genie-02 + OmniPicker grippers

| Resolution | 8–9 cameras, up to 1920 × 1536

| Frame rate | 30 fps; G2 home stereo 60 fps

| Format | LeRobot v3.0

| Views | 8–9 video streams

| Capture modes | Whole-body and dual-arm

| G1 cameras | 8 views: main, three head fisheye, two back fisheye, two wrist

| G2 cameras | 8 home views; 9 commercial views including depth

| Video encoding | H.265

| State | 9 groups covering 14 arm joints, currents, end-effector poses, grippers, head, waist and body pose

| Action | 7 groups covering joints, end-effectors, grippers, head, waist and body velocity

| Delivery | Parquet + MP4 + metadata; native format also included for selected tasks

| Capture variation | Resolution and frame rate vary by camera view. Depth is included in G2 commercial-scene packages only.

INSIDE THE DATASET

## Package and coverage

One independent LeRobot v3.0 dataset per task, organized by robot, scene, location and task.

### Annotation

Scene → location → task hierarchy, with an individual task description for each recording.

### Coverage

53 episodes across 11 robot/task groups in homes, a hotel and a library.

REPRESENTATIVE WORK

## Sample tasks

- Sofa tidying
- Tabletop rubbish removal
- Sandwich preparation
- T-shirt folding
- Hotel changing-area organization
- Library shelf organization
- Packing stationery

EXPLORE THE CATALOGUE

## Find the right data for your model.

[Next: Galbot Robot Data ↗](https://actionable.ai/datasets/R02.html)[All eight collections ↗](https://actionable.ai/#catalogue)