Hoshipuother

LIBERO Logical State and Action Trajectories

LIBERO robot manipulation trajectories augmented with per-frame logical state and logical action annotations across five task suites. Contains 6425 demonstrations with synchronized logical predicates and actions aligned to raw sensor frames.

Downloads0
Episodes6425

Why This Matters for Physical AI

This dataset enables training of imitation learning models and logical reasoning systems by providing aligned raw sensorimotor trajectories with high-level logical state and action annotations, supporting research on action/state abstraction and language grounding in robotic manipulation.

Technical Profile

Modalities
rgbproprioceptionlanguage
Action Space
joint_positions
Environment
simulation
Task Types
manipulationpick_and_place
Episodes
6425
Data Format
HDF5
Annotation Types
language_instructionsaction_labelslogical_state_predicates
License
other
Part of the LIBERO 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