mteb/mrtidy

该数据集包含多种语言的语料库、查询相关性和查询内容，涵盖了阿拉伯语、孟加拉语、英语、芬兰语、印尼语、日语、韩语、俄语、斯瓦希里语、泰卢固语和泰语。每个语言的配置分为corpus、qrels和queries三个部分，分别表示语料库、查询相关性和查询内容。每个部分包含特征如_id、title、text等，并且数据集被分割为train、dev和test三个部分，每个部分都有对应的文件大小和样本数量。

This dataset contains corpora, query relevance, and query content in multiple languages, including Arabic, Bengali, English, Finnish, Indonesian, Japanese, Korean, Russian, Swahili, Telugu, and Thai. Each language configuration is divided into three parts: corpus, qrels, and queries, representing the corpus, query relevance, and query content, respectively. Each part includes features such as _id, title, text, etc., and the dataset is split into train, dev, and test sections, each with corresponding file sizes and sample counts.