mlfoundations-dev/filtered_pdf_flow

该数据集包含文档相关的信息，如文档的URL、文件名、是否成功处理等。还包括页面数量、页码、页面字节数等文档属性。此外，数据集还包含问题、选项、答案以及一些处理后的字段。训练集共有12938个样本，数据集主要用于文本提取和问题回答相关的任务。

The dataset contains document-related information such as the document URL, filename, and whether it was processed successfully. It also includes document attributes like page count, page number, and page byte size. In addition, the dataset contains questions, options, answers, and some post-processed fields. The training set consists of 12,938 samples, and the dataset is primarily used for tasks related to text extraction and question answering.