ttslearn.wavenet.gen.gen_waveform

ttslearn.wavenet.gen.gen_waveform(device, labels, logf0_vuv, wavenet_model, wavenet_in_scaler, binary_dict, numeric_dict, tqdm=<class 'tqdm.std.tqdm'>)[source]

Generate waveform from WaveNet.

Parameters
  • device (torch.device) – torch.device to use.

  • labels (nnmnkwii.io.hts.HTSLabel) – full context labels.

  • logf0_vuv (torch.Tensor) – Log-f0 and V/UV flag.

  • wavenet_model (nn.Module) – Trained WaveNet.

  • wavenet_in_scaler (sklearn.preprocessing.StandardScaler) – StandardScaler for WaveNet input.

  • binary_dict (dict) – Dictionary of binary features.

  • numeric_dict (dict) – Dictionary of numeric features.

  • tqdm (tqdm) – tqdm progress bar.

Returns

Generated waveform.

Return type

numpy.ndarray