Use LSTM to learn feature mapping from corrupted to clean speech Please see another repo CNN_denoise. This project simply replaces the CNN with an LSTM. The pipeline is exactly the same.
weedwind/LSTM_denoise
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|