LDKP3K

该数据集包含了大约10万篇带有关键词标记的长文档，其中一个小版本的数据集包含了2万训练样本、3,413个验证样本以及3,339个测试样本。该数据集是通过将KP20K语料库映射到S2ORC生成的。在规模上，它包括2万训练样本、3,413个验证样本以及3,339个测试样本，其任务是进行关键词提取。

This dataset comprises approximately 100,000 long documents annotated with keywords. A small-scale variant of this dataset includes 20,000 training samples, 3,413 validation samples, and 3,339 test samples. This dataset is generated by mapping the KP20K corpus to S2ORC. The small-scale variant follows the aforementioned sample splits, and its core task is keyword extraction.