mteb/XGlueWPRReranking

XGLUE是一个用于评估跨语言预训练模型在跨语言自然语言理解和生成方面性能的新基准数据集。它包含11个任务，涵盖了19种语言，适合作为跨语言模型的基准测试。数据集包含各种配置，包括德语、英语、法语、意大利语、葡萄牙语、西班牙语和中文等语言的语料库、查询、相关性和顶级排名数据。每个配置都有验证和测试两个数据集，并提供相应的下载大小和示例数量。XGLUE数据集源自forresty/xglue数据集，并使用MTEB库进行评估。

XGLUE is a new benchmark dataset designed for evaluating the performance of cross-lingual pre-trained models in natural language understanding and generation tasks. It consists of 11 tasks spanning 19 languages, making it suitable for benchmarking cross-lingual models. The dataset includes various configurations for languages such as German, English, French, Italian, Portuguese, Spanish, and Chinese, encompassing corpus, queries, relevance, and top-ranked data. Each configuration features validation and test datasets with specified download sizes and example counts. XGLUE is derived from the forresty/xglue dataset and is evaluated using the MTEB library.