Medix-AI/coqui-tts
Medix-AI
/
coqui-tts
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix synthesize.py

This commit is contained in:
erogol 2020-02-19 18:24:02 +01:00
parent f8ebf6abcd
commit e540a54959
1 changed files with 7 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
synthesize.py
View File

@ -25,14 +25,16 @@ def tts(model,
t_1 = time.time() t_1 = time.time()
use_vocoder_model = vocoder_model is not None use_vocoder_model = vocoder_model is not None
waveform, alignment, _, postnet_output, stop_tokens = synthesis( waveform, alignment, _, postnet_output, stop_tokens = synthesis(
model, text, C, use_cuda, ap, speaker_id, False, model, text, C, use_cuda, ap, speaker_id, style_wav=False,
C.enable_eos_bos_chars) truncated=False, enable_eos_bos_chars=C.enable_eos_bos_chars,
use_griffin_lim=(not use_vocoder_model), do_trim_silence=True)
if C.model == "Tacotron" and use_vocoder_model: if C.model == "Tacotron" and use_vocoder_model:
postnet_output = ap.out_linear_to_mel(postnet_output.T).T postnet_output = ap.out_linear_to_mel(postnet_output.T).T
# correct if there is a scale difference b/w two models # correct if there is a scale difference b/w two models
if use_vocoder_model:
postnet_output = ap._denormalize(postnet_output) postnet_output = ap._denormalize(postnet_output)
postnet_output = ap_vocoder._normalize(postnet_output) postnet_output = ap_vocoder._normalize(postnet_output)
if use_vocoder_model:
vocoder_input = torch.FloatTensor(postnet_output.T).unsqueeze(0) vocoder_input = torch.FloatTensor(postnet_output.T).unsqueeze(0)
waveform = vocoder_model.generate( waveform = vocoder_model.generate(
vocoder_input.cuda() if use_cuda else vocoder_input, vocoder_input.cuda() if use_cuda else vocoder_input,
@ -58,7 +60,7 @@ if __name__ == "__main__":
parser.add_argument( parser.add_argument(
'out_path', 'out_path',
type=str, type=str,
help='Path to save final wav file.', help='Path to save final wav file. Wav file will be names as the text given.',
) )
parser.add_argument('--use_cuda', parser.add_argument('--use_cuda',
type=bool, type=bool,