Closed chkwon1 closed 5 years ago
Well, the model is created by https://github.com/fatchord/WaveRNN and not me, but I suspect that there are probably too few model parameters for that. Still, might be worth a try.
close #6
Will be solved once we have #8
I modified bits=9 to 16 in wavernn/preprocess.py and model.py. By this modification your code is working. Do you think the synthesized speech quality can be improved?