Home · Open data · Teleoperation · Robot video

GR00T Teleop GR1

Roughly 88 hours across 22,209 teleoperated Fourier GR-1 humanoid episodes at 1280x800, 20 fps, used to post-train NVIDIA's DreamDojo. Non-commercial.

Published by NVIDIA, 2026-02

Open data
Non-commercial only

CC BY-NC 4.0

PriceFree

The publisher's licence restricts use to research and other non-commercial purposes.

Get dataset Opens huggingface.co in a new tab
From a terminal
huggingface-cli download --repo-type dataset nvidia/PhysicalAI-Robotics-GR00T-Teleop-GR1 --local-dir ./PhysicalAI-Robotics-GR00T-Teleop-GR1

Public, ungated, no token needed. Three top-level folders (GR1_robot, DreamDojo-HV_Eval, EgoDex_Eval) that can be pulled separately with --include. Total size is not published; 6.36M frames of 1280x800 h264 video implies a large download, check with huggingface-cli before starting. Licence CC BY-NC 4.0, non-commercial. Usage instructions live in https://github.com/NVIDIA/DreamDojo.

Specification, as published

Hours (derived)
88 h
Episodes (stated)
22,209 episodes
Scale
7.55M rows (HF viewer, default subset); GR1_robot meta: total_episodes 22209, total_frames 6362293
Task
Pick & place, Other
Environment
Indoor – lab
Modality
RGB, Proprioception, Actions, Language
Capture
Egocentric
In frame
Robot hands
Captured in
North America
Embodiment
Fourier GR-1 T1 humanoid (robot_type GR1T1, embodiment_tag gr1_unified)
Frame rate
20 fps
Resolution
1280x800
Formats
LeRobot v2, Parquet, MP4, JSON
Hosted on
Hugging Face

Figures are the publisher's own. Kinetic Blocks has not measured this dataset. Derived from the Hub metadata for GR1_robot: 6,362,293 frames at 20 fps. The small evaluation folders are not counted.

About this dataset

The real-robot post-training data behind NVIDIA's DreamDojo world model, recorded on a Fourier GR-1 T1 humanoid by teleoperation at 20 Hz. The main GR1_robot folder holds 22,209 episodes and about 6.36 million frames with a single 1280x800 ego view, state and action vectors, rewards and human coarse and fine action annotations in LeRobot v2 format. Two smaller folders, DreamDojo-HV_Eval and EgoDex_Eval, contain robot replications of scenes from human video datasets used for evaluation. The card is tagged CC BY-NC 4.0 and carries almost no dataset documentation of its own.

Publisher's own task names: GR1_robot post-training set, DreamDojo-HV_Eval, EgoDex_Eval, human video replication tasks.

What Kinetic Blocks did, and did not do

Indexed and linked from the publisher. Not hosted, verified or graded by Kinetic Blocks. The publisher's terms govern. The licence shown is the one stated on the publisher's page, linked above. Nothing on this page is a claim by Kinetic Blocks.

Want the verified ones too?

Request access to the marketplace, where every paid listing is verified against its files and graded before it goes on sale.