Install python packages $ pip install torch kaldi-python-io librosa soundfile Download pre-trained models - d-vector extractor model - uis-rnn model Prepare data file - wav data - vad data run example $ run.sd_batch.sh