Skip to content
/ HCRN Public
forked from aispeech-lab/HCRN

PyTorch implementation of HCRN described in our paper "Towards Modeling Auditory Restoration in Noisy Environments"

Notifications You must be signed in to change notification settings

vyouman/HCRN

 
 

Repository files navigation

HCRN

PyTorch implementation of Hourglass-shaped Convolutional Recurrent Network (HCRN) described in our paper "Towards Modeling Auditory Restoration in Noisy Environments"

Requirements

Pytorch=0.4.1
resampy
soundfile
pysepq
librosa
pystoi
lera

Data Preparation

Use the scripts in preprocess to process the WSJ, ESC-50 and Audioset datasets. Please modify the dataset paths in the scripts.

Train and test

For training: python3 Train.py

For testing: pytho3 Test.py

Please modify specify processed dataset paths in config.yaml.

Contact

If you have any questions, please feel free to contact me at [email protected]

Citations

@article{huang2021Towards, title={Towards Modeling Auditory Restoration in Noisy Environments}, author={Huang, Yating and Hao, Yunzhe and Xu, Jiaming and Xu, Bo} journal={Proc. IJCNN 2021} year={2021} }

About

PyTorch implementation of HCRN described in our paper "Towards Modeling Auditory Restoration in Noisy Environments"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%