WikiMT++

WikiMT++是由中央音乐学院和复旦大学合作创建的音乐数据集，包含1010首精心挑选的乐谱，采用ABC记谱法。该数据集不仅继承了WikiMT的特性，还增加了客观属性（如专辑、歌词、视频）和主观情感属性（12种情感形容词），增强了其在音乐信息检索、条件音乐生成、自动作曲和情感分类等领域的应用。数据集的创建过程中，通过Python脚本筛选出完整描述的乐谱，并利用CLaMP模型进行属性校正，确保数据的准确性和完整性。WikiMT++适用于多种音乐分析任务，如音乐信息检索、音乐生成、歌词自动生成、流派识别和情感分类，为音乐研究和分析提供了丰富的资源。

WikiMT++ is a music dataset jointly developed by the Central Conservatory of Music and Fudan University, which contains 1010 carefully selected musical scores in ABC notation. Building on the characteristics of the original WikiMT dataset, WikiMT++ additionally incorporates objective attributes (e.g., album, lyrics, and video) and subjective emotional attributes (12 types of emotional adjectives), thereby expanding its application scenarios in domains including music information retrieval, conditional music generation, automatic composition, and emotion classification. During the dataset construction process, Python scripts were utilized to filter out fully annotated musical scores, and the CLaMP model was employed for attribute calibration to guarantee the accuracy and integrity of the dataset. WikiMT++ is applicable to a wide range of music analysis tasks, such as music information retrieval, music generation, automatic lyrics generation, genre recognition, and emotion classification, serving as a valuable and abundant resource for music research and analysis.