search

keithito / tacotron

A TensorFlow implementation of Google's Tacotron speech synthesis with pre-trained model (unofficial)
MIT License
2.96k stars 957 forks source link

About pretrained model(ljspeech) #197

Closed sunnnnnnnny closed 6 years ago

sunnnnnnnny commented 6 years ago

Hello, I want to ask you how many steps are trained based on ljspeech. I downloaded the extracted file name to model.ckpt.data-00000-of-00001. It is very strange and I hope to get your reply.

keithito commented 6 years ago

The model was trained for 877K steps, though there was not much quality improvement after step 250K.