SpaRRTa-Lego: Real-World Spatial-Relation Benchmark with Toy Minifigures

SpaRRTa-Lego is the real-world counterpart of a synthetic spatial-relation benchmark. Scenes are photographed with toy minifigures and everyday objects for evaluating Visual Foundation Models. The dataset is associated with a paper (arXiv:2601.11729) and code, and was last updated on 2026-06-25.

Use Cases

• Sim-to-real evaluation of spatial-relation capabilities based on the described real-world photographs.
• 4-way classification (Front/Back/etc.) based on the task description for spatial understanding.
• Benchmarking Visual Foundation Models on a real-world visual reasoning task.

Strengths

• Designed as a real-world counterpart to a synthetic benchmark, enabling sim-to-real evaluation.
• Associated with a published paper (arXiv:2601.11729) and open-source code, providing research context.

Limitations

• Description metadata is limited; actual data quality, size, and column structure require manual inspection after download.
• Row count is unknown, which may limit suitability assessment for large-scale training.

Provenance

Source

huggingface user turhancan97

Collection Method

Scenes are photographed with toy minifigures and everyday objects.

Freshness

Last updated 2026-06-25 15:01:09; freshness should be verified.