Lingbot Depth Subset: 13,149 Samples for Depth Perception from Three Sub-Collections

13,149 samples organized into 10,207 groups across three distinct sub-collections named RobbyReal, RobbyVla, and RobbySim. This dataset is hosted on Hugging Face by the organization Voxel51 and was last updated on July 14, 2026. The data appears to be structured for use with the FiftyOne computer vision toolkit.

Use Cases

• Training depth estimation models based on the three named sub-collections (RobbyReal, RobbyVla, RobbySim).
• Benchmarking robotic vision systems using the 10,207 sample groups.
• Developing multimodal perception pipelines based on the dataset's likely combination of visual and depth data.
• Analyzing performance differences between simulated (RobbySim) and real-world (RobbyReal, RobbyVla) data sources.

Strengths

• Contains 13,149 individual samples, providing a substantial base for model development.
• Data is organized into 10,207 groups, which may indicate structured scenes or sequences.
• Comprises three distinct sub-collections, allowing for comparative analysis across data sources.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale training.
• The description metadata is limited; actual data quality and content require manual inspection after download.

Provenance

Source

Voxel51 via Hugging Face.

Collection Method

Likely collected or generated for robotics and computer vision applications, involving real and simulated environments.

Freshness

Last updated 2026-07-14 20:40:52; freshness should be verified.