auDeep: Unsupervised Learning of Representations from Audio with Deep\n Recurrent Neural Networks

Abstract

auDeep is a Python toolkit for deep unsupervised representation learning from\nacoustic data. It is based on a recurrent sequence to sequence autoencoder\napproach which can learn representations of time series data by taking into\naccount their temporal dynamics. We provide an extensive command line interface\nin addition to a Python API for users and developers, both of which are\ncomprehensively documented and publicly available at\nhttps://github.com/auDeep/auDeep. Experimental results indicate that auDeep\nfeatures are competitive with state-of-the art audio classification.\n