TTS model loading failed

[Dong-mel]

Hi friends,
When I load the model using model_path, a language attribute error occurs, but it works fine when I load it using model_name.

model_path loader:

# model_name = 'tts_models/fr/css10/vits'
model_path = "test0830-September-02-2024_11+44AM-dbf1a08a\best_model.pth"
model_config = "test0830-September-02-2024_11+44AM-dbf1a08a\config.json"

# Get device
device = "cuda" if torch.cuda.is_available() else "cpu"
# Init TTS
tts = TTS(model_path=model_path, config_path=model_config).to(device)

Error log:

Traceback (most recent call last):
  File "python3.9/site-packages/TTS/api.py", line 332, in tts_to_file
    self._check_arguments(speaker=speaker, language=language, speaker_wav=speaker_wav, **kwargs)
  File "python3.9/site-packages/TTS/api.py", line 228, in _check_arguments
    if self.is_multi_lingual and language is None:
  File "python3.9/site-packages/torch/nn/modules/module.py", line 1729, in __getattr__
    raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'")
AttributeError: 'TTS' object has no attribute 'is_multi_lingual'

When I was debugging, I found that self.synthesizer.tts_model contained "language_manager" and had a value.

Has anyone encountered this error? Or have any suggestions? Thanks!

[eginhard]

This bug (duplicate of #3449) is fixed in our fork, available via pip install coqui-tts

[Dong-mel]

Thank you very much, it works fine now!