r9y9 / deepvoice3_pytorch

PyTorch implementation of convolutional neural networks-based text-to-speech synthesis models
https://r9y9.github.io/deepvoice3_pytorch/
Other
1.97k stars 485 forks source link

Training Multi-Speaker Model on vctk got bad results! #197

Open the-final opened 4 years ago

the-final commented 4 years ago

Hi, r9y9, I trained 600k steps of multi-speaker model on the preprocessed vctk, but the performance is not as good as the model you released. I see that the number of training steps of your model is 300K + 300K , and the learning rate of your model is 5.7e-5. Can you explain the training process? Thanks! @r9y9