Closed edwardhdlu closed 5 years ago
Implementation of the WaveNet architecture for speech synthesis.
Local conditioning is supported, so we can generate audio autoregressively by conditioning on mel spectrograms generated from Tacotron2. Fast inference is not yet supported.
Implementation of the WaveNet architecture for speech synthesis.
Local conditioning is supported, so we can generate audio autoregressively by conditioning on mel spectrograms generated from Tacotron2. Fast inference is not yet supported.