Live

New this week23 datasets listed, 645,000+ hSee what is listed

Home · Open data · Manipulation · Robot video

UnifoLM R1 Dex1 Dataset

Roughly 0.8 hours and 169 episodes of Unitree R1 humanoid two-arm teleoperation with Dex1 grippers: towel folding, pouring, setting plates; Apache 2.0.

Published by Unitree Robotics, 2026-10

Fisheye head camera view of a Unitree R1 with two grippers holding a blue towel on an office desk
Open data

Frame from the UnifoLM R1 Dex1 Dataset (R1_Dex1_Fold_Towel), Unitree Robotics, Apache 2.0.

Commercial use OK

Apache 2.0

PriceFree

The publisher's licence permits commercial use and redistribution, with attribution. Hub card tag apache-2.0 on all five repos, applied by Unitree's own organisation; the card bodies carry no licence sentence of their own.

Get dataset Opens huggingface.co in a new tab
From a terminal
hf download unitreerobotics/R1_Dex1_Fold_Towel --repo-type dataset --local-dir ./R1_Dex1_Fold_Towel

Five public repos listed in the collection, about 4.2 GB together; the command pulls the towel folding one.

Specification, as published

Hours (derived)
0.8 h
Episodes (stated)
169 episodes
Task
Laundry & folding, Pick & place, Tidying & put-away, Cooking & food prep
Environment
Indoor – office
Modality
RGB, Proprioception, Actions
Capture
Egocentric
In frame
Robot hands
Captured in
Asia
Embodiment
Unitree R1 humanoid with Dex1 parallel grippers
Frame rate
30 fps
Resolution
640x480
Formats
LeRobot v3.0, Parquet, MP4
Size
about 4.2 GB on the Hub across 5 repos
Hosted on
Hugging Face

Figures are the publisher's own. Kinetic Blocks has not measured this dataset. Derived from the meta/info.json of all five repos: 89,372 frames at 30 fps. Frames are timesteps and are not summed across the four camera views.

About this dataset

A first, small release on Unitree's R1 humanoid: upper-body two-arm teleoperation with Dex1 parallel grippers in an office setting. Five task repos cover folding a towel, pouring popcorn between bowls, putting pens into a drawer, setting plates and tidying building blocks. Each episode has two head cameras and two wrist cameras at 640x480 and 30 fps, with 16-dimensional arm and gripper state and action vectors.

Publisher's own task names: fold the towel, pour popcorn between bowls, put pens into drawer, set plates, tidy building blocks.

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 verified listings are checked against their files and graded A+ to F, and your money waits until the data holds up.