yufan

You may use two channels of audio data. Our model only supports single channel audio.

you should add link to kaldi utils and steps. ln -s /kaldi/utils/ ./

Our baseline model is based on Espnet and Kaldi. I suggest you learn how to install and use Espnet and Kaldi first. https://github.com/kaldi-asr/kaldi https://github.com/espnet/espnet

You can download our AliMeeting data at https://www.openslr.org/119 .