Explore high-quality datasets for your AI and machine learning projects.
The dataset comprises approximately 1,700 music excerpts in .mp3 format, each lasting 270–300 seconds and sampled at 22 kHz. The excerpts are taken from NetEase Cloud Music and are labelled with 16 genre categories. The dataset is divided into a Raw Subset and an Eval Subset, each providing different audio features and annotations. It was created to foster AI research in the music industry and was mainly collected and annotated by students. The dataset is intended for audio‑classification tasks and supports multilingual use.
