irds/mmarco_v2_pt_dev

`mmarco/v2/pt/dev`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101,093个查询（queries）和59,273个相关性评估（qrels）。文档（docs）需要从另一个数据集`irds/mmarco_v2_pt`中获取。使用示例展示了如何通过Hugging Face的datasets库加载数据。

The `mmarco/v2/pt/dev` dataset is provided by the ir-datasets package and is primarily utilized for text retrieval tasks. It contains 101,093 queries and 59,273 relevance judgments (qrels). The document (docs) corpus needs to be obtained from another dataset `irds/mmarco_v2_pt`. Usage examples demonstrate how to load the dataset via Hugging Face's datasets library.

原始信息汇总

数据集概述

数据集名称

mmarco/v2/pt/dev

来源

数据集来源于ir-datasets包，原始数据集为irds/mmarco_v2_pt。

数据内容

• queries: 查询（即主题），数量为101,093。
• qrels: 相关性评估，数量为59,273。
• docs: 文档数据，需从irds/mmarco_v2_pt获取。

使用示例

python from datasets import load_dataset

queries = load_dataset(irds/mmarco_v2_pt_dev, queries) for record in queries: record # {query_id: ..., text: ...}

qrels = load_dataset(irds/mmarco_v2_pt_dev, qrels) for record in qrels: record # {query_id: ..., doc_id: ..., relevance: ..., iteration: ...}

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} }