DCS-Net

This is the implementation of "DCS-Net: Deep Complex Subtractive Neural Network for Monaural Speech Enhancement"

Setup

1. Edit VOICEBANK_ROOT in config.py to be the root of the dataset

2. Install the dependencies

   $ pip install -r requirements.txt

3. Adjust hyperparameters in config.py

Training

{DCS-Net: dcs, DRS-Net: drs, DC-Net: dc, DR-Net: dr}

$ python train.py [NETWORK]

Testing

{DCS-Net: dcs, DRS-Net: drs, DC-Net: dc, DR-Net: dr}

$ python test.py [NETWORK]