COVIDROP is the first numerical reasoning based machine reading comprehension dataset in the Vietnamese language for the COVID-19 pandemic domain. Consists of 6594 question-answers pairs from 841 news articles which are collected from the four reliable and reputable Vietnamese online news websites are Thanhien, VNExpress, Zing News and BaoMoi. The questions and answers are manually annotated by members of the DS@UIT Laboratory. For more details, please refer to our original thesis in Vietnamese here.
Notice: This dataset is published for research purposes only. The dataset is not intended for commercial use.