when I use the official vocos and bigvgan to infer the model, I found the results really hard to listen, it seems that I used a wrong vocoder to reconstruct waveform.
· Sign up or log in to comment
