Multilingual Document Classification Corpus (MLDoc)

MLDoc是由Facebook AI Research创建的多语言文档分类数据集，基于Reuters Corpus Volume 2构建，包含487,000篇新闻故事，涵盖英语、法语、西班牙语、意大利语、德语、俄语、中文和日语八种语言。数据集通过均匀抽样确保每种语言和类别的覆盖均衡，分为训练、开发和测试集，旨在评估跨语言文档分类的性能。MLDoc的应用领域广泛，主要用于解决跨语言文档分类问题，推动相关领域的研究发展。

MLDoc is a multilingual document classification dataset developed by Facebook AI Research, constructed based on the Reuters Corpus Volume 2. It contains 487,000 news stories across 8 languages, namely English, French, Spanish, Italian, German, Russian, Chinese, and Japanese. Uniform sampling is adopted to ensure balanced coverage for each language and category. The dataset is split into training, development, and test subsets, with its core purpose being to evaluate the performance of cross-lingual document classification tasks. It has a wide range of application scenarios, mainly used to address cross-lingual document classification problems and promote research and development in related fields.