Rafael Valle

What do you mean by 3% loss? Loss should be between 0 and + infinity. Did you train on a single vctk speaker or all speakers?

Take a look at this [link](http://sergeiturukin.com/2017/03/02/wavenet.html)

The topic has been discussed on several issues, including 112 below. Please close this issue to avoid repetition. https://github.com/ibab/tensorflow-wavenet/issues/112

Let us know once you find out what causes it!

@feiyuhong can you share link to some code? I'm interested in porting it to pytorch!

Probably inside util/html.py

Please also use the code below to replace the clean method on the SEGAN class because it is extremely slow. def clean(self, x): """ clean a utterance x x: numpy...

This link below is a fork of the code that works with tensorflow 1.3.0 https://github.com/lordet01/segan One can also follow @santi-pdp recommendation and use tensorflow 0.12 with cuDNN 5.1.

I think they have that information on the arxiv paper On Nov 11, 2017 12:52 AM, "Gaofeng Cheng" wrote: > @rafaelvalle how long will it take to > finish the...

@linan2 look at my Oct 16 message.