mulliganMIT

real-routing-d2-c02-dagger-mulligan-no-cf

A training dataset variant from the Mulligan project for real-world cable routing task (real-routing-d2) using dagger learning without counterfactual augmentation. Part of an ablation study with 50 episodes collected from multiple camera viewpoints.

Downloads323
Episodes50

Why This Matters for Physical AI

This dataset contributes to physical AI research by providing real-world demonstrations of cable routing manipulation tasks with multi-view camera observations, enabling training of learned policies for precise object manipulation in constrained environments.

Technical Profile

Modalities
rgb
Environment
lab
Task Types
manipulationrouting
Episodes
50
Data Format
parquet
Annotation Types
language_instructionsreward_labels
License
MIT
Part of the Mulligan 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