search

andabi / deep-voice-conversion

Deep neural networks for voice conversion (voice style transfer) in Tensorflow
MIT License
3.92k stars 843 forks source link

Some questions about the performance and the amount of the data #25

Open JiaYK opened 6 years ago

JiaYK commented 6 years ago

Hello, The result of what I use your demo to train is not very good, so it was not as seem as yours. At the seem time, I used some Chinese database to train, the effect was still not good. So I wonder to know if the parameter of train2 which you shown on the web page is the default parameter. What about the amount of data and how many is it?

jiqizaisikao commented 6 years ago

Hi,Im using some Chinese database to train too,which data do you use to train net1 Contact Q:3617 28654