Update app.py

app.py

@@ -25,21 +25,24 @@ def get_text(text, hps, is_phoneme):
 
 
 def create_tts_fn(model, hps, speaker_ids):
- def tts_fn(text, speaker, speed, is_symbol):
+ def tts_fn(text, speaker, speed, is_phoneme):
  if limitation:
- text_len = len(re.sub("\[([A-Z]{2})\]", "", text))
- max_len = 150
- if is_symbol:
+ text_len = len(text)
+ max_len = 100
+ if is_phoneme:
  max_len *= 3
+ else:
+ if len(hps.data.text_cleaners) > 0 and hps.data.text_cleaners[0] == "zh_ja_mixture_cleaners":
+ text_len = len(re.sub("(\[ZH\]|\[JA\])", "", text))
  if text_len > max_len:
  return "Error: Text is too long", None
 
  speaker_id = speaker_ids[speaker]
- stn_tst = get_text(text, hps, is_symbol)
+ stn_tst = get_text(text, hps, is_phoneme)
  with no_grad():
- x_tst = stn_tst.unsqueeze(0).to(device)
- x_tst_lengths = LongTensor([stn_tst.size(0)]).to(device)
- sid = LongTensor([speaker_id]).to(device)
+ x_tst = stn_tst.unsqueeze(0)
+ x_tst_lengths = LongTensor([stn_tst.size(0)])
+ sid = LongTensor([speaker_id])
  audio = model.infer(x_tst, x_tst_lengths, sid=sid, noise_scale=.667, noise_scale_w=0.8,
  length_scale=1.0 / speed)[0][0, 0].data.cpu().float().numpy()
  del stn_tst, x_tst, x_tst_lengths, sid