daniel-kwon/therpay_benchmarking

该数据集包含已摄入和已摘要的两种类型的文档数据。在已摄入的数据中，包含文档ID、文档文本、文档文件名和文件大小等元数据。在已摘要的数据中，除了包含已摄入数据的内容外，还包括原始文档摘要、生成的摘要文本和用于生成摘要的模型名称。数据集提供了一个训练集分割。

The dataset consists of two configurations: ingested and summarized. Both configurations include document ID, document text, document filename, and document metadata (file size). The summarized configuration also includes the raw document summary, generated document summary, and the name of the summarization model. The dataset is split into a training set, with different sizes for each configuration.