NetherlandsForensicInstitute/squad-nl-v2.0

SQuAD-NL v2.0数据集是一个用于句子相似性和问答系统任务的数据集，特别适用于评估句子嵌入模型。该数据集包含问题、上下文、分数、ID、标题和答案等特征，其中分数用于指示问题是否在上下文中找到答案。数据集是从SQuAD和XQuAD英文数据集翻译而来，特别是通过Google Translate进行翻译，测试集还进行了人工校对。数据集分为训练集、开发集和测试集，建议仅使用测试集来测试荷兰语句子嵌入模型。

The SQuAD-NL v2.0 dataset is a Dutch translation of the original SQuAD and XQuAD datasets, specifically designed for sentence similarity tasks in Sentence Transformers. The dataset includes features such as question, context, score, id, title, and answers. The score column is added to indicate whether the question has an answer in the context, with 1.0 indicating an answer and 0.0 indicating no answer. The dataset is divided into train, dev, and test splits, with the test split recommended for evaluating Dutch sentence embedding models. The original license for the dataset is CC BY-SA 4.0.