hbredin/VoiceActivityDetection-PyanNet-DIHARD

Voice activity detection trained on DIHARD III development set . Uses pyannote.audio 2.0 (which is still in development) The simplest way of getting voice activity detection results is to use the pretrained pipeline . If you need more control (e.g. to lower the detection threshold for better recall) the model can be loaded like that . The model can also be loaded as a model or an annotated version of the pipeline . For example, the model is loaded with the model and the pipeline is called Inference . The pipeline can be used to run the pipeline or the pipeline to get the results . The audio files can be provided as a 'waveform' numpy,