Dataset · airoa-org
YUBI robotic video dataset for CoRL 2026 (part 3/5)
airoa-org/yubi-corl2026-umi-arena-03-2k0
Third of five parts of the large-scale stereo video dataset by AIRoA for robot task learning.
What's inside
This is the third of five parts of the YUBI release for The UMI Arena competition at CoRL 2026, prepared by AIRoA. The full release includes over 2,000 hours of 30 Hz video (over 217M frames and 961K episodes) across 805 tasks with English descriptions. Recordings are captured from two cameras (left and right) at 480x640 resolution under a restricted license.
What you can do with it
- Training Vision-Language-Action (VLA) models and robot control policies
- Training binocular perception algorithms for robotic manipulators
- Preparing solutions for The UMI Arena track at CoRL 2026
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 22.1K
- Likes
- 0
- Access
- on request on Hugging Face
How to load it
from datasets import load_dataset
ds = load_dataset("airoa-org/yubi-corl2026-umi-arena-03-2k0")