MIR-1K

该数据集包含了用于训练旋律提取模型的中文流行歌曲集合，它与MedleyDB结合使用，共同为训练旋律提取模型提供支持。该数据集规模包含1000首歌曲，主要任务是对歌唱旋律的提取。

This dataset comprises a corpus of Chinese pop songs intended for training melody extraction models. It is used in conjunction with MedleyDB to jointly support the training of such models. This dataset includes 1000 songs, and its primary task focuses on singing melody extraction.