huzheyuanapache-2.0

3lego WHOLE-EPISODE sessions: val_base

Dual-arm YAM robot lego block-sorting dataset reconverted to whole-episode format for in-context learning, with multi-cycle recordings under fixed block-to-tray permutations.

Downloads0
Episodes4
Hours0.07

Why This Matters for Physical AI

Whole-episode trajectories with latent task permutations enable training in-context learning models for robotic manipulation under variable environmental configurations.

Technical Profile

Modalities
rgbproprioceptionjoint_state
Robot Embodiments
Dual-YAM
Action Space
end_effector_delta
Environment
lab
Task Types
manipulationpick_and_placesorting
Episodes
4
Total Hours
0.07
Data Format
LeRobot
Annotation Types
language_instructionsaction_labels
License
apache-2.0
Part of the 3lego family

Access

Need custom rgb data?

Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets