CATIE-AQ/piaf_fr_prompt_qa
收藏资源简介:
piaf_fr_prompt_qa是法语提示数据集(DFP)的一个子集,包含387,408行数据,适用于问答任务。原始数据来自PIAF数据集,并通过FrenchQA数据集中的SQUAD 2.0格式的问题进行了增强。通过应用一系列提示,构建了输入和目标列,使其格式与Muennighoff等人创建的xP3数据集一致。数据集包含42个提示,分为训练集,没有验证集和测试集。
piaf_fr_prompt_qa is a subset of the French Prompt Dataset (DFP), containing 387,408 rows of data and designed for question answering tasks. Its raw data originates from the PIAF dataset, and it is augmented with questions formatted in SQUAD 2.0 from the FrenchQA dataset. By applying a series of prompts, the input and target columns were constructed to align with the format of the xP3 dataset created by Muennighoff et al. The dataset includes 42 prompts and is split exclusively into a training set, with no validation or test sets provided.
piaf_fr_prompt_qa 数据集概述
基本信息
- 语言: 法语
- 许可证: MIT
- 数据规模: 100K<n<1M
- 任务类别: 问答
- 标签: DFP, french prompts
- 标注创建者: found
- 语言创建者: found
- 多语言性: 单语种
- 源数据集: etalab-ia/piaf
数据集详情
- 名称: piaf_fr_prompt_qa
- 概述: 该数据集是 Dataset of French Prompts (DFP) 的一个子集,包含 387,408 条数据,适用于问答任务。原始数据来自 PIAF 数据集,并增加了 FrenchQA 数据集中的问题,格式为 SQUAD 2.0。通过应用一系列提示,构建了输入和目标列,以获得与 xP3 数据集相同的格式。
提示列表
- 数量: 42 个提示
- 逻辑: 提示采用直陈式,包括 tutoiement 和 vouvoiement 形式。
数据集划分
- 训练集: 387,408 条样本
- 验证集: 无
- 测试集: 无
使用方法
python from datasets import load_dataset dataset = load_dataset("CATIE-AQ/piaf_fr_prompt_qa")
引用
原始数据
@InProceedings{keraron-EtAl:2020:LREC, author = {Keraron, Rachel and Lancrenon, Guillaume and Bras, Mathilde and Allary, Frédéric and Moyse, Gilles and Scialom, Thomas and Soriano-Morales, Edmundo-Pavel and Staiano, Jacopo}, title = {Project PIAF: Building a Native French Question-Answering Dataset}, booktitle = {Proceedings of The 12th Language Resources and Evaluation Conference}, month = {May}, year = {2020}, address = {Marseille, France}, publisher = {European Language Resources Association}, pages = {5483--5492}, url = {https://www.aclweb.org/anthology/2020.lrec-1.673} }
本数据集
@misc {centre_aquitain_des_technologies_de_linformation_et_electroniques_2023,
author = { {Centre Aquitain des Technologies de lInformation et Electroniques} },
title = { DFP (Revision 1d24c09) },
year = 2023,
url = { https://huggingface.co/datasets/CATIE-AQ/DFP },
doi = { 10.57967/hf/1200 },
publisher = { Hugging Face }
}




