Pytorch port of Google Research's VGGish model used for extracting audio features. - View it on GitHub