mteb/MIRACLRetrieval

MIRACLRetrieval是一个多语言检索数据集，专注于跨18种不同语言的搜索。它包括各种元数据，如注释创建者、支持的语种、许可证、任务类别和数据集配置。每个配置都包括特征、分割、下载大小和数据集大小。此外，文件还提供了如何使用MTEB框架在此数据集上评估模型的说明，并鼓励在使用数据集时进行引用。

MIRACLRetrieval is a multilingual retrieval dataset that focuses on search across 18 different languages. It includes various metadata such as annotations creators, supported languages, license, task categories, and dataset configurations. Each configuration includes features, splits, download size, and dataset size. Additionally, the file provides instructions on how to evaluate models on this dataset using the MTEB framework and encourages citation if the dataset is used.