Discogs-EffNet (ONNX) — unmodified mirror
Unmodified mirror of the Discogs-EffNet music style model (discogs-effnet-bsdynamic-1.onnx)
and its Genre Discogs400 labels (genre_discogs400-discogs-effnet-1.json), created by the
Music Technology Group (MTG), Universitat Pompeu Fabra, and published with Essentia.
No changes were made to the files.
Used by the spotify-genres Spicetify extension to infer genres locally.
Credit
- Model author: Pablo Alonso-Jiménez (MTG), with Xavier Serra and Dmitry Bogdanov.
- Copyright © Universitat Pompeu Fabra.
- Original files and documentation: https://essentia.upf.edu/models.html
License
MTG's documentation licenses its models under
CC BY-NC-SA 4.0 (non-commercial use only);
a proprietary license for commercial use is available from MTG on request
(https://www.upf.edu/web/mtg/contact). MTG's own LICENSE file, included here verbatim,
carries the CC BY-NC-ND 4.0 legal text; an unmodified, non-commercial mirror is permitted
under either. The files are provided as-is, without warranties of any kind
(see Section 5 of the license).
Citation
The model (MTG's metadata for these files):
@inproceedings{alonso2022music,
title={Music Representation Learning Based on Editorial Metadata from Discogs},
author={Alonso-Jim{\'e}nez, Pablo and Serra, Xavier and Bogdanov, Dmitry},
booktitle={Conference of the International Society for Music Information Retrieval (ISMIR)},
year={2022}
}
Essentia's TensorFlow models, as MTG asks for research use:
@inproceedings{alonso2020tensorflow,
title={Tensorflow Audio Models in {Essentia}},
author={Alonso-Jim{\'e}nez, Pablo and Bogdanov, Dmitry and Pons, Jordi and Serra, Xavier},
booktitle={International Conference on Acoustics, Speech and Signal Processing ({ICASSP})},
year={2020}
}