Eval Inference Test: Robot Arm Action Data for Evaluation

A dataset created using the LeRobot framework, likely containing action data for robot control. The dataset's meta information indicates it includes a 6-dimensional action feature representing joint positions for a robotic arm. It was uploaded by hobby321 to Hugging Face and was last updated on June 4, 2026.

Use Cases

• Evaluate robot action prediction models based on the 6-dimensional joint position data.
• Benchmark reinforcement learning policies for robotic manipulation using the provided action sequences.
• Test simulation-to-reality transfer for robotic arms using the structured action data.

Strengths

• Action data is structured with a defined shape (6 dimensions) and data type (float32).
• The dataset was created using the LeRobot framework, suggesting a standardized generation process.
• Metadata specifies a codebase version (v3.0) and frame rate (30 fps) for context.

Limitations

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

Provenance

Source

hobby321 via Hugging Face.

Collection Method

Created using the LeRobot framework.

Freshness

Last updated 2026-06-04 08:37:46; freshness should be verified.