search

breizhn / DTLN

Tensorflow 2.x implementation of the DTLN real time speech denoising model. With TF-lite, ONNX and real-time audio processing support.
MIT License
567 stars 160 forks source link

There is no any code with Training data preparation? #45

Open shenbuguanni opened 3 years ago

echogaurav commented 2 years ago

You can check the data preparation code here. If you already have a mapped set of clean an noisy audio files in wav format then you can even use that, just ensure that audio lengths are consistent across the dataset.

StuartIanNaylor commented 2 years ago

I just ran it clone @breizhn repo of https://github.com/breizhn/DNS-Challenge and run the splt and runs fine