Hunterls

rewind-branching-k2-d4

A branching search dataset containing 106 searches across 14 manipulation tasks, with 6,432 nodes capturing states, images, and outcomes of action chunk proposals in simulated kitchen, study, and office environments.

Downloads0
Episodes106

Why This Matters for Physical AI

This dataset enables training of action selection and outcome prediction models by providing counterfactual branches showing what would have happened with alternative action chunks, supporting research in search-based planning and policy learning.

Technical Profile

Modalities
rgbproprioception
Action Space
action_chunks
Environment
simulation
Task Types
manipulationpick_and_placeobject_placement
Episodes
106
Data Format
parquet
Annotation Types
action_labelsoutcome_labels
Part of the RoboPRO family

Access

Need custom rgb data?

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

Request a Sample Pack

Related Datasets