Dataset Card for MTG Jamendo Dataset

Dataset Summary

MTG-Jamendo Dataset, a new open dataset for music auto-tagging. It is built using music available at Jamendo under Creative Commons licenses and tags provided by content uploaders. The dataset contains over 55,000 full audio tracks with 195 tags from genre, instrument, and mood/theme categories. We provide elaborated data splits for researchers and report the performance of a simple baseline approach on five different sets of tags: genre, instrument, mood/theme, top-50, and overall.

Dataset structure

Data Fields

id: an integer containing the id of the track
artist_id: an integer containing the id of the artist
album_id: an integer containing the id of the album
duration_in_sec: duration of the track as a float
genres: list of strings, describing genres the track is assigned to
instruments: list of strings for the main instruments of the track
moods: list of strings, describing the moods the track is assigned to
audio: audio of the track

Data Splits

This dataset has 2 balanced splits: train (90%) and validation (10%)

Licensing Information

This dataset version 1.0.0 is released under the Apache-2.0 License.

Citation Information

@conference {bogdanov2019mtg,
    author = "Bogdanov, Dmitry and Won, Minz and Tovstogan, Philip and Porter, Alastair and Serra, Xavier",
    title = "The MTG-Jamendo Dataset for Automatic Music Tagging",
    booktitle = "Machine Learning for Music Discovery Workshop, International Conference on Machine Learning (ICML 2019)",
    year = "2019",
    address = "Long Beach, CA, United States",
    url = "http://hdl.handle.net/10230/42015"
}