Skip to content

Dataset · ccmusic-database

NetEase music tracks across 16 genres

ccmusic-database/music_genre

Around 1,700 audio recordings (metadata indicates a 10K-100K scale) from NetEase with genre labels for music classification tasks.

What's inside

The dataset contains around 1,700 .mp3 audio clips ranging from 270 to 300 seconds long with a 22,050 Hz sampling rate (general repository metadata indicates 10K to 100K records). Sourced from NetEase Music in Chinese and English, the 16 genre labels are pulled directly from the website without manual annotators. Image classification is also listed among the tasks.

What you can do with it

  • Training and testing models for automatic music genre classification from audio
  • Training classifiers on spectrograms generated from audio tracks
Tasks
Audio classification, Image classification
Languages
Chinese, English
Size
10K–100K rows
Downloads this month
446
Likes
66

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("ccmusic-database/music_genre")