kiyam/ddro-pseudo-queries
收藏官方服务：
资源简介：
ddro-pseudo-queries数据集包含了用于训练DDRO模型的生成伪查询，这些查询是基于DocTTTTTQuery方法生成的，旨在扩展生成文档检索的训练数据。数据集包含两个文件，分别针对MS MARCO和Natural Questions数据集，每个文件中包含了与文档ID对应的伪查询，每个文档对应10个伪查询。
The ddro-pseudo-queries dataset contains generated pseudo queries used for training DDRO models, created based on the DocTTTTTQuery approach to expand the training data for generative document retrieval. The dataset includes two files, one for the MS MARCO dataset and one for the Natural Questions dataset, each containing pseudo queries mapped to corresponding document IDs, with 10 pseudo queries per document.
提供机构：
kiyam
