commul/ud_genre

UD Genre Labels 2.17 是一个为通用依存树库提供的句子级体裁标注数据集。这些标签通过自举管道生成，并非权威的黄金标注。数据集包含2,198,282个句子，覆盖18种不同体裁，标注方法包括bootstrap-inferred、single-genre-treebank和virtual-split。它用于文本分类任务，支持多语言，适用于体裁分类和句子分类研究。

UD Genre Labels 2.17 is a dataset providing derived sentence-level genre annotations for Universal Dependencies treebanks. These labels are produced by the bootstrapping pipeline and are not authoritative gold annotations. The dataset includes 2,198,282 sentences across 18 genres, with methods such as bootstrap-inferred, single-genre-treebank, and virtual-split. It is designed for text-classification tasks, supports multilingual applications, and is relevant for genre-classification and sentence-classification research.