TensorFlow Audio Models in Essentia
Audio and Speech Processing
2020-03-18 v1 Machine Learning
Sound
Abstract
Essentia is a reference open-source C++/Python library for audio and music analysis. In this work, we present a set of algorithms that employ TensorFlow in Essentia, allow predictions with pre-trained deep learning models, and are designed to offer flexibility of use, easy extensibility, and real-time inference. To show the potential of this new interface with TensorFlow, we provide a number of pre-trained state-of-the-art music tagging and classification CNN models. We run an extensive evaluation of the developed models. In particular, we assess the generalization capabilities in a cross-collection evaluation utilizing both external tag datasets as well as manual annotations tailored to the taxonomies of our models.
Keywords
Cite
@article{arxiv.2003.07393,
title = {TensorFlow Audio Models in Essentia},
author = {Pablo Alonso-Jiménez and Dmitry Bogdanov and Jordi Pons and Xavier Serra},
journal= {arXiv preprint arXiv:2003.07393},
year = {2020}
}