This repository contains code for training and evaluating convolutional neural networks (CNNs) for speech classification tasks. The models in this repository have been tested for the language identification in speech. In addition to conventional CNNs, this repository features speech classifiers that implements domain adaptation using adversarial training, which are also known as domain-adversarial neural networks (DANNs).
The code is fairly documented and the vectorization logic, as well as the code for the models, should be useful for other speech technology tasks. If you use our code and encounter problems, please create an issue or contact the first author.
If you use our code for another publication, please cite our paper as
@inproceedings{abdullah20_interspeech,
author={Badr M. Abdullah and Tania Avgustinova and Bernd Möbius and Dietrich Klakow},
title={{Cross-Domain Adaptation of Spoken Language Identification for Related Languages: The Curious Case of Slavic Languages}},
year=2020,
booktitle={Proc. Interspeech 2020},
pages={477--481},
doi={10.21437/Interspeech.2020-2930}
}