GlowTTS - Single Speaker ,

[md844108]

I train GlowTTS model using same dataset as given in Recipes on ljspeech with 10 epocs, model is trained but when we infrencing the model, output come as audio file but audio file not contain proper word. So can you suggest me why if this low epocs then tell me minimum amount of epocs required to perfect output?