MEDIC

该数据集包含71,198张图片，分为七大类别：地震、火灾、洪水、飓风、山体滑坡、其他灾害以及非灾害。在数据划分上，训练集占70%，验证集占20%，测试集占10%。经过筛选后，图片数量从71,198张减少至38,094张。在规模上，筛选后有38,094张图片。该数据集的任务是多模态灾害分类。

This dataset contains 71,198 images, which are classified into seven major categories: earthquakes, wildfires, floods, hurricanes, landslides, other disasters, and non-disaster samples. Regarding data splitting, 70% of the total data is assigned to the training set, 20% to the validation set, and the remaining 10% to the test set. After filtering, the number of images was reduced from 71,198 to 38,094, which is the final size of the dataset. The task of this dataset is multimodal disaster classification.