ThinkGeoBench Question-4 split contains a set of agricultural remote sensing benchmark questions. The dataset includes image assets and depth data for tasks like apple size estimation. It was uploaded by author yz686868 to Hugging Face on May 4, 2026.
Use Cases
- Benchmarking agricultural object detection models based on referenced image assets.
- Evaluating depth estimation algorithms for fruit size analysis based on AppleSizeEstimate .npy files.
- Developing question-answering systems for geospatial agricultural data based on the original question file.
- Studying agricultural taxonomy and task templates based on the provided taxonomy summary.
Strengths
- Includes normalized metadata for Hugging Face Dataset Viewer via metadata.jsonl.
- Contains specialized depth data assets for apple size estimation.
- Provides original source questions and a taxonomy summary for task definition.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- Hugging Face dataset uploaded by yz686868.
- Freshness
- Last updated 2026-05-04 08:54:21.