Wikipedia Paragraphs
收藏arXiv2025-09-30 收录
下载链接:
https://dl.fbaipublicfiles.com/dpr/wikipedia_split/psgs_w100.tsv.gz
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含了来自维基百科的2100万个段落,它被用作MATTER模型的一个非结构化知识库。其规模达到了2100万个段落,适用于问答任务。
This dataset consists of 21 million paragraphs sourced from Wikipedia, serving as an unstructured knowledge base for the MATTER model. With a corpus size of 21 million paragraphs, it is well-suited for question answering tasks.



