linmou

Thanks for your warmly reply. Any vocoder recommend? I want to inverse fbank features to audios.

I only achieve 63 in f1c. Maybe there is something wrong with my data preprocess progress.