LeWAM Community Dataset: 321 Robotics Datasets for Manipulation Tasks

LeWAM Community Dataset is a large-scale collection of 321 robotics datasets contributed by the LeRobot community. It is derived from Community Dataset v2 and converted to LeRobotDataset v3.0, serving as an artifact from the LeWAM project. The dataset focuses on primary tasks like manipulation, pick & place, and sorting, featuring robot types such as SO-100 and 101.

Use Cases

• Training robot learning models for manipulation tasks based on the described primary task focus.
• Developing algorithms for robotic pick-and-place operations based on the dataset's stated task categories.
• Benchmarking sorting algorithms for robotics based on the included task type.
• Studying robot behavior across different robot types (e.g., SO-100/101) based on the description.

Strengths

• Contains 321 distinct datasets, indicating a broad collection.
• Features a weighted average frame rate of 30.0 FPS for video data.
• Covers specific primary robotics tasks: manipulation, pick & place, and sorting.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and total data size are unknown, which may limit suitability assessment.
• License information is not provided in the available metadata.

Provenance

Source

LeRobot community, aggregated by author ehalicki on Hugging Face.

Collection Method

Community-contributed, derived from Community Dataset v2 and converted to LeRobotDataset v3.0.

Time Range

null

Freshness

Last updated 2026-03-31 14:55:14; freshness should be verified.

Geography

null