fatchord / WaveRNN

WaveRNN Vocoder + TTS

Home Page:https://fatchord.github.io/model_outputs/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

gen_tacotron got noise voice

lucasjinreal opened this issue · comments

this is pretrained model tacotron voice result

2_batchedTrue_180k.zip

it's seems all noise data but not voice

@jinfagang Is that sample produced by the quick_start.py script?

@fatchord No, simply generated by tacotron model.

@jinfagang The pretrained models are automatically loaded in only for the quick_start.py script. The gen_tacotron.py file is for your own models (although you can overwrite the weights in the checkpoints folder with the pretrained weights if you wish).

I am also getting the same issue