oxe-augecc-by-4.0

berkeley_autolab_ur5_train_500_600

Name: berkeley_autolab_ur5_train_500_600
Creator: oxe-auge
License: cc-by-4.0
Keywords: rgb, proprioception, language, manipulation, lab

An augmented robotics dataset containing 100 episodes with multi-robot cross-embodiment annotations across 9 different robot platforms, augmented from the original Berkeley UR5 dataset using OXE-Aug framework.

Downloads233
Episodes100

Why This Matters for Physical AI

This dataset enables cross-embodiment policy learning by providing multi-robot augmentations of manipulation demonstrations, allowing models to generalize manipulation skills across diverse robot morphologies.

Technical Profile

Modalities: rgbproprioceptionlanguage
Robot Embodiments: google_robotjacokinova3kuka_iiwapandasawyerwidowXxarm7
Action Space: joint_positions
Environment: lab
Task Types: manipulation
Episodes: 100
Data Format: parquet
Annotation Types: language_instructionsaction_labels
License: cc-by-4.0

Part of the OXE-Aug family

Community Signals

Top 50% by downloads

Access

View on HuggingFace Read Paper

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

berkeley_autolab_ur5_train_400_500

An augmented robotics dataset containing 100 episodes with multi-robot vision and proprioceptive data, featuring cross-embodiment trajectory augmentation for 9 different robot platforms including Franka Panda, UR5, and others.

rgbproprioception

276 downloads2025cc-by-4.0

berkeley_autolab_ur5_train_800_896

An augmented robotics dataset containing 96 episodes with cross-embodiment transformations across 9 different robot platforms (google_robot, jaco, kinova3, kuka_iiwa, panda, sawyer, widowX, xarm7) generated using OXE-Aug v2.1.

rgbproprioception

219 downloadscc-by-4.0

berkeley_autolab_ur5_train_300_400

An augmented robotics dataset containing 100 episodes with cross-embodiment image transformations for 9 different robot platforms (Google Robot, Jaco, Kinova3, KUKA IIWA, Franka Panda, Sawyer, WidowX, xArm7). The dataset includes joint positions, end-effector poses, and natural language instructions with 5 FPS frame rate.

rgbproprioceptionlanguage

189 downloads2025cc-by-4.0

berkeley_autolab_ur5_train_0_100

An augmented robotics dataset containing 100 episodes of manipulation tasks with robot-specific image augmentations for 9 different robot embodiments, providing cross-embodiment policy learning data with joint positions, end-effector poses, and natural language instructions.

rgbproprioceptionlanguage

178 downloadscc-by-4.0

berkeley_autolab_ur5_train_600_700

An augmented robotics dataset containing 100 episodes with cross-embodiment visual and proprioceptive observations across 9 robot platforms, augmented from the Berkeley UR5 demonstration dataset.

rgbproprioception

171 downloads2025cc-by-4.0

berkeley_autolab_ur5_train_700_800

An augmented robotics dataset containing 100 episodes with multi-robot cross-embodiment image augmentations for 9 different robot platforms. The dataset provides RGB observations, joint positions, end-effector poses, and natural language instructions for manipulation tasks.

rgbproprioceptionlanguage

168 downloads2025cc-by-4.0