A multimodal benchmark for evaluating large vision-language models on extreme weather event analysis tasks.
ObsCrisis-Bench contains 4,202 VQA samples across 127 extreme weather events in 8 disaster categories, covering 61 countries. Each sample combines satellite multispectral imagery (AMSU-A, HIRS, MHS sensors) with optional weather station data, and requires models to perform risk assessment, type classification, timing… See the full description on the dataset page:
https://huggingface.co/datasets/YYQ898/Obshazard-bench.
