Dataset · nvidia
NVIDIA GR1 Robot Teleoperation Simulation Dataset
nvidia/PhysicalAI-Robotics-GR00T-Teleop-Sim
1,000 simulated teleoperation trajectories of the GR1 humanoid robot for tabletop manipulation training.
What's inside
This dataset from the NVIDIA GEAR team contains 1,000 teleoperation trajectories of the upper body of the GR1 robot with Fourier hands in a simulator. It includes RGB video streams from a virtual camera and tabular data (1M to 10M records). The dataset is open for non-commercial use.
What you can do with it
- Training VLA models and robot manipulation control policies in simulation
- Sim-to-real transfer of tabletop object manipulation skills to humanoid robots
- Size
- 1M–10M rows
- Downloads this month
- 22.5K
- Likes
- 20
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("nvidia/PhysicalAI-Robotics-GR00T-Teleop-Sim")