LIBERO LeRobot V3: Robot Manipulation Benchmark for Lifelong Learning

nvidia/LIBERO_LeRobot_v3 is a LeRobotDataset v3.0 conversion of the LIBERO benchmark for lifelong robot learning. The dataset packages task suites with Parquet state/action data, MP4 video observations, and LeRobot metadata. It was last updated on 2026-05-31.

Use Cases

• Training robot manipulation policies based on state/action data.
• Studying knowledge transfer across tasks based on the lifelong learning benchmark design.
• Developing vision-language-action models based on MP4 video observations and language conditioning.
• Benchmarking robot learning algorithms based on the structured LIBERO task suites.

Strengths

• Includes multiple data modalities: Parquet files for state/action data and MP4 files for video observations.
• Designed for a specific research domain: lifelong robot learning and knowledge transfer.
• Converted to a standardized format (LeRobotDataset v3.0) for compatibility.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

NVIDIA

Collection Method

Conversion of the LIBERO robot manipulation benchmark into a LeRobot-compatible format.

Freshness

Last updated 2026-05-31 19:01:13; freshness should be verified.