rishikksh20 / HiFiplusplus-pytorch

HiFi++: a Unified Framework for Neural Vocoding, Bandwidth Extension and Speech Enhancement
MIT License
151 stars 19 forks source link

which is better for TTS, hifigan vs hifi++? #5

Open Liujingxiu23 opened 2 years ago

Liujingxiu23 commented 2 years ago

which is better for TTS, hifigan vs hifi++?

rishikksh20 commented 2 years ago

right now I guess iSTFTNet is best compared to both hifigan and hifi++. Currently, I am struggling to implement hifi++ architecture correctly. The author didn't share much info regarding training hyperparameters thats why this hifi++ implementation underperform compared to original hifigan

Liujingxiu23 commented 2 years ago

Thank very much for you reply, I will try to train iSTFTNet for TTS.