Classical Poetry Subject Clustering Benchmark
数据集描述
A classical poetry subject clustering benchmark for evaluating embedding models' performance on unsupervised poetry clustering tasks. 每条记录包含一首诗歌及其所属的题材聚类标签。
本数据集按照 8:2 的比例拆分为训练集和测试集,并保持 subject_code 的分布。
id: 诗歌唯一标识符
dynasty: 诗歌朝代
author: 诗歌作者
title: 诗歌标题
type: 诗歌体裁(如五言律诗)
text: 诗歌正文内容
subject:… See the full description on the dataset page:
https://huggingface.co/datasets/PoetryMTEB/ClassicalPoetrySubjectClustering.