mtg/music-classifiers

Public

Transfer learning models for music classification by genres, moods, and instrumentation

806 runs

Latest version

Predictions run on Nvidia T4 GPU hardware.

This demo runs transfer learning classifiers trained on various public and in-house MTG datasets using different audio embeddings.

Source models used for embeddings

MusiCNN. A musically motivated CNN with two variants trained on the Million Song Dataset and the MagnaTagATune.
VGGish. A large VGG variant trained on a preliminary version of the AudioSet Dataset.

Our models consist of single-hidden-layer MLPs trained on the considered embeddings.

These models are part of Essentia Models made by MTG-UPF and are publicly available under CC by-nc-sa and commercial license.