deu05232/promptriever-RQ1-only_q_version
收藏资源简介:
该数据集是一个用于训练信息检索或问答系统模型的结构化数据集,包含训练分割(train),共有211,853个示例。数据特征包括:query_id(查询ID)、query(查询文本)、positive_passages(正相关文档列表,含docid、text、title等字段)、negative_passages(负相关文档列表,含docid、text、title等字段)、only_instruction(仅指令文本)、only_query(仅查询文本)、has_instruction(布尔值指示是否包含指令)和new_negatives(序列字段)。数据集下载大小为1,034,159,145字节,总大小为1,761,730,895字节。
This dataset is a structured dataset for training information retrieval or question-answering system models, containing a training split (train) with 211,853 examples. Features include: query_id, query, positive_passages (a list of relevant documents with fields such as docid, text, title), negative_passages (a list of non-relevant documents with fields such as docid, text, title), only_instruction, only_query, has_instruction (a boolean indicating whether instructions are included), and new_negatives (a sequence field). The download size is 1,034,159,145 bytes, and the total dataset size is 1,761,730,895 bytes.



