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

Don't print computed phonemes 

It causes noise in logs
This commit is contained in:
Eren Gölge 2021-07-13 10:59:05 +02:00
parent d63a6bb690
commit 8fffd4e813
3 changed files with 2 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
TTS/tts/datasets/TTSDataset.py
View File

@ -244,6 +244,7 @@ class TTSDataset(Dataset):
# TODO: find a better fix # TODO: find a better fix
return self.load_data(self.rescue_item_idx) return self.load_data(self.rescue_item_idx)
pitch = None
if self.compute_f0: if self.compute_f0:
pitch = self._load_or_compute_pitch(self.ap, wav_file, self.f0_cache_path) pitch = self._load_or_compute_pitch(self.ap, wav_file, self.f0_cache_path)

2
TTS/tts/layers/losses.py
View File

@ -687,7 +687,7 @@ class FastPitchLoss(nn.Module):
spec_loss = self.spec_loss(decoder_output, decoder_target, decoder_output_lens) spec_loss = self.spec_loss(decoder_output, decoder_target, decoder_output_lens)
ssim_loss = self.ssim(decoder_output, decoder_target, decoder_output_lens) ssim_loss = self.ssim(decoder_output, decoder_target, decoder_output_lens)
dur_loss = self.dur_loss(dur_output[:, : ,None], dur_target[:, :, None], input_lens) dur_loss = self.dur_loss(dur_output[:, :, None], dur_target[:, :, None], input_lens)
pitch_loss = self.pitch_loss(pitch_output.transpose(1, 2), pitch_target.transpose(1, 2), input_lens) pitch_loss = self.pitch_loss(pitch_output.transpose(1, 2), pitch_target.transpose(1, 2), input_lens)
loss = ( loss = (
self.spec_loss_alpha * spec_loss self.spec_loss_alpha * spec_loss

3
TTS/tts/utils/text/__init__.py
View File

@ -45,12 +45,10 @@ def text2phone(text, language, use_espeak_phonemes=False):
# TO REVIEW : How to have a good implementation for this? # TO REVIEW : How to have a good implementation for this?
if language == "zh-CN": if language == "zh-CN":
ph = chinese_text_to_phonemes(text) ph = chinese_text_to_phonemes(text)
print(" > Phonemes: {}".format(ph))
return ph return ph
if language == "ja-jp": if language == "ja-jp":
ph = japanese_text_to_phonemes(text) ph = japanese_text_to_phonemes(text)
print(" > Phonemes: {}".format(ph))
return ph return ph
if gruut.is_language_supported(language): if gruut.is_language_supported(language):
@ -80,7 +78,6 @@ def text2phone(text, language, use_espeak_phonemes=False):
# Fix a few phonemes # Fix a few phonemes
ph = ph.translate(GRUUT_TRANS_TABLE) ph = ph.translate(GRUUT_TRANS_TABLE)
return ph return ph
raise ValueError(f" [!] Language {language} is not supported for phonemization.") raise ValueError(f" [!] Language {language} is not supported for phonemization.")