LivUMI Grip Sample: Dual-Arm Robot Imitation Learning Data

A dataset for imitation and robot learning, formatted for the LeRobot v2.1 framework. It contains episodes of data from the LivUMI dual-arm robot, including multiple camera feeds, depth information, proprioceptive state, and time indexing. The dataset was created by livsynrobotics and last updated on 2026-04 08.

Use Cases

• Train imitation learning policies for robotic grasping based on multimodal sensor data.
• Benchmark robot learning algorithms using episodes with synchronized camera, depth, and proprioceptive state.
• Develop visuomotor control models based on the included multiple camera perspectives.
• Analyze temporal action sequences in robotic tasks using the time-indexed episode data.

Strengths

• Data is structured in a standardized LeRobot-compatible v2.1 format, easing integration with that ecosystem.
• Includes multiple synchronized data modalities per episode, such as cameras, depth, and proprioception.
• Last update timestamp is explicitly provided (2026-04-08 08:29:33).

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Row count, file formats, and column-level documentation are absent, which may limit suitability assessment.
• Data may reflect bias inherent to the specific LivUMI robot platform and its operational environment.

Provenance

Source

livsynrobotics

Collection Method

Likely collected from the LivUMI dual-arm robot platform during operation.

Time Range

null

Freshness

Last updated 2026-04-08 08:29:33.

Geography

null