irds/mmarco_it_train

`mmarco/it/train`数据集由ir-datasets包提供，主要用于文本检索任务。该数据集包含808,731个查询（queries）、532,761个相关性评估（qrels）和39,780,811个文档对（docpairs）。文档（docs）部分需要使用`irds/mmarco_it`数据集。

The `mmarco/it/train` dataset, provided by the ir-datasets package, is primarily intended for text retrieval tasks. It contains 808,731 queries, 532,761 relevance assessments (qrels), and 39,780,811 document pairs (docpairs). The document (docs) component requires the use of the `irds/mmarco_it` dataset.

原始信息汇总

数据集概述

数据集名称

mmarco/it/train

数据集来源

由ir-datasets提供。

数据集内容

• queries（查询，即主题）：数量为808,731
• qrels（相关性评估）：数量为532,761
• docpairs：数量为39,780,811

数据集使用

• 对于docs数据，使用irds/mmarco_it数据集。

示例代码

python from datasets import load_dataset

queries = load_dataset(irds/mmarco_it_train, queries) for record in queries: record # {query_id: ..., text: ...}

qrels = load_dataset(irds/mmarco_it_train, qrels) for record in qrels: record # {query_id: ..., doc_id: ..., relevance: ..., iteration: ...}

docpairs = load_dataset(irds/mmarco_it_train, docpairs) for record in docpairs: record # {query_id: ..., doc_id_a: ..., doc_id_b: ...}

引用信息

@article{Bonifacio2021MMarco, title={{mMARCO}: A Multilingual Version of {MS MARCO} Passage Ranking Dataset}, author={Luiz Henrique Bonifacio and Israel Campiotti and Roberto Lotufo and Rodrigo Nogueira}, year={2021}, journal={arXiv:2108.13897} }