Hiwonder-robot/cube_to_plate

SO-ARM101 Cube2Plate Dataset

Dataset Description

This dataset contains demonstration data collected for the Cube2Plate task using the SO-ARM101 robotic arm.

The task requires the robot arm to pick up a cube and place it onto a plate. The dataset is intended for training and evaluating imitation learning policies with the LeRobot framework.

Task

Pick up a cube and place it onto a plate.

Dataset Details

• Dataset name: SO-ARM101 Cube2Plate Dataset
• Dataset repository: Hiwonder-robot/cube_to_plate
• Robot: SO-ARM101
• Task: Cube2Plate
• Framework: LeRobot
• Collection method: Teleoperation
• License: Apache-2.0

Data Format

The dataset follows the LeRobot dataset format.

The dataset includes robot demonstration trajectories and visual observations. The available image observations include:

• observation.images.front
• observation.images.handeye

The dataset may also include robot states, actions, timestamps, episode indices, and frame indices depending on the LeRobot recording configuration.

Data Collection

The data was collected by operating the SO-ARM101 robotic arm to complete the Cube2Plate task.

Each episode records a demonstration trajectory where the robot moves a cube and places it onto a plate.

Intended Use

This dataset is intended for:

• Training SO-ARM101 manipulation policies
• Imitation learning experiments
• Behavior cloning
• LeRobot policy training and evaluation
• Cube-to-plate robotic manipulation research

How to Use

Use the following dataset repository ID with LeRobot:

Hiwonder-robot/cube_to_plate

Example:

from lerobot.common.datasets.lerobot_dataset import LeRobotDataset

dataset = LeRobotDataset("Hiwonder-robot/cube_to_plate")
frame = dataset[0]

print(frame.keys())

Related Model

A policy model trained with this dataset is available as:

Hiwonder-robot/Cube2Plate

Limitations

• This dataset is designed for the Cube2Plate task only.
• Performance may depend on camera position, lighting conditions, object placement, and robot calibration.
• The dataset may not generalize well to different objects, environments, or robot configurations.
• Real-world robot deployment should be tested carefully to avoid collision or hardware damage.

Citation

If you use this dataset, please cite or reference this Hugging Face dataset repository:

Hiwonder-robot/cube_to_plate

More Information

This dataset was collected and uploaded for use with the LeRobot framework.