CLARU
mulliganMIT

real-marker-d2-c01-dagger-mixed

Part of the Mulligan robotics dataset collection, containing real robot manipulation trajectories collected for the real-marker-d2 task with mixed sources and dagger methodology.

Downloads239
Episodes115

Why This Matters for Physical AI

This dataset contributes real robot manipulation trajectories with multi-view observations to support training and evaluation of embodied AI policies through the Mulligan benchmark framework.

Technical Profile

Modalities
rgb
Environment
lab
Task Types
manipulation
Episodes
115
Data Format
parquet
Annotation Types
reward_labels
License
MIT
Part of the Mulligan family

Access

View on HuggingFace

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

real-routing-d2-c09-dagger-mixed

Part of the Mulligan robotics dataset collection, containing real robot routing task recordings with mixed sources from model round R5. Includes 200 episodes with multi-camera observations collected at 15 FPS.

rgb
627 downloadsOct 2026mit

real-routing-d2-c08-dagger-mixed

Part of the Mulligan project, this dataset contains raw robotic routing task recordings with mixed sources, including trajectory data, episode metadata, and multi-camera observations collected at 15 FPS.

rgb
399 downloadsOct 2026MIT

real-routing-d2-c00-teleop-mixed

A real robot manipulation dataset from the Mulligan project containing 250 episodes of teleop-collected mixed routing tasks with multi-camera observations.

rgb
365 downloadsOct 2026MIT

real-routing-d2-c07-dagger-mixed

A robotics dataset part of the Mulligan project containing real routing task demonstrations with mixed data sources collected across multiple cameras. Comprises 103 episodes with 51,569 frames recorded at 15 FPS for policy learning and evaluation.

rgb
320 downloadsOct 2026MIT

real-routing-d2-legacy-r03-eval-b01

An evaluation dataset of 100 episodes from the real-routing-d2 task, part of the Mulligan robotics benchmark collection. Contains multi-camera observations and trajectory data for policy evaluation.

rgb
295 downloadsOct 2026MIT

real-marker-d2-c00-teleop-mixed

A real robot manipulation dataset collected via teleoperation containing 250 episodes of marker-based task demonstrations. Part of the Mulligan project, providing raw teleoperated collection data with multi-camera observations.

rgb
287 downloadsOct 2026MIT