r-three/tiny-xquad

xquad是一个多语言阅读理解数据集，包含阿拉伯语、德语、希腊语、英语、西班牙语、印地语、罗马尼亚语、俄语、泰语、土耳其语、越南语和中文等12种语言。数据集中的每个样本都包含一个上下文字段(context)、一个问题字段(question)和一个答案字段(answers)，答案字段包括答案文本和答案在上下文中的起始位置。该数据集适用于阅读理解模型的训练和评估，每种语言版本都提供了验证集，共包含119个样本。

xquad is a multilingual reading comprehension dataset including Arabic, German, Greek, English, Spanish, Hindi, Romanian, Russian, Thai, Turkish, Vietnamese, and Chinese. Each sample in the dataset contains a context field, a question field, and an answers field which includes the answer text and the starting position of the answer in the context. This dataset is suitable for training and evaluating reading comprehension models, with each language version providing a validation set containing 119 samples.