dataset · Global

remote-sensing-VQA-benchmark

Hugging Face dataset AdaptLLM/remote-sensing-VQA-benchmark

Claim this listing

Represent remote-sensing-VQA-benchmark? Send a claim for review so we can discuss listing updates. Claims do not automatically grant ownership or verification.

Explore paid sponsorship →

What's in it

Data offered docs-verified: remote-sensing image category questions (VQA) · image-question-answer pairs over 256x256 px aerial scenes · scene categories such as airport, bare land, beach, bridge, forest, port, railway
Coverage llm-inferred: Global
Resolution docs-verified: 256 x 256 px image tiles (ground sample distance not stated)
Formats docs-verified: arrow
Access llm-inferred: Public Hugging Face dataset page; licence not stated in docs

Useful for

Use cases llm-inferred: Benchmark a vision-language model's scene classification of aerial imagery before using it for site-photo or survey triage · Evaluate flood-scene recognition using the FloodNet subset as a test of model robustness to flood imagery · Compare candidate VLMs on remote-sensing VQA to inform model selection for GIS image-labelling workflows
Audience llm-inferred: GIS analysts · remote sensing specialists · ML engineers evaluating vision-language models · hydrologists and flood analysts (FloodNet subset)

Limitations

Constraints llm-inferred: Test split only (15k rows); no training split listed · English-only questions · Images are 256x256 px with no ground resolution stated, so not directly usable for measurement · Subset row counts (CLRS 15k, FloodNet 4.5k, NWPU-Captions 3.15k, UC_Merced 2.1k) do not obviously reconcile with the stated 15k test split · Source imagery comes from third-party datasets; licence not stated in docs or ENTRY · Not UK-specific; scene categories are generic rather than UK habitat or flood-risk classes · Benchmark for VQA evaluation rather than an operational data feed; no update frequency stated

Licence: Unknown — check provider terms
Authentication: Unknown — review source docs
GitHub stars: Unknown
Last commit: Unknown
Last successful page check: 14/10/2026
Page health: healthy
Package: Not recorded
Verification: Not independently verified

Provenance & evidence

Verified 2026-10-15 via HF API: exists; 201 downloads/30d, 4 likes.

Metadata observed 14/10/2026 · research sweep: manual-2026-10-15 (Hugging Face datasets)

https://huggingface.co/datasets/AdaptLLM/remote-sensing-VQA-benchmark