Simulation GR1 Tabletop Task 1K Dataset

Dataset Description:

The PhysicalAI-Robotics-GR00T-Teleop-GR1 dataset consists of 1000 teleoperation trajectories in simulation using the GR1 robot with upper body control. The simulation setup mimics tabletop manipulation tasks and uses RGB observations with a virtual camera. The robot is equipped with simulated Fourier hands.

This dataset is ready for non-commercial use.

Dataset Owner(s):

NVIDIA GEAR

Dataset Creation Date:

June 1, 2025

License/Terms of Use:

This dataset is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).

Intended Usage:

Researchers, Academics, Open-Source Community: AI-driven robotics research and algorithm development in simulation environments. Developers: Integrate and customize AI policies for tabletop tasks. Startups & Companies: Non-commercial benchmarking and prototyping.

Dataset Characterization

Data Collection Method
[Human Teleoperation in Simulation]

Labeling Method
[Human]

Dataset Format

HDF5 format and LeRobot-style dataset

Modality:

• Observation: 44 dim of vectorized state (joint positions of upper body + waist)
• Action: 44 dim of vectorized action (joint positions of upper body + waist)
• Video: RGB video, 256x256 resolution (cropped and padded; original resolution is 51x320), 20 fps
• Language Instruction Example:
  • "pick up the bottled water, place it into the cabinet and close the cabinet"
  • "pick the cupcake from the cutting board and place it in the basket"

Dataset Quantification

• 1000 teleoperation trajectories for each of the 24 tabletop tasks
• Total data storage (HDF5 format): ~14GB
• Total data storage (LeRobot-style dataset): ~39GB

Download the dataset

huggingface-cli download \
    --repo-type dataset nvidia/PhysicalAI-Robotics-GR00T-Teleop-Sim \
    --local-dir ./datasets/