Offline usage?

[George0828Zhang]

Tried to load model locally using this:

def load_vocos(path, device):
    config_path = os.path.join(path, 'config.yaml')
    model_path = os.path.join(path, 'pytorch_model.bin')
    model = Vocos.from_hparams(config_path)
    state_dict = torch.load(model_path, map_location="cpu")
    if isinstance(model.feature_extractor, EncodecFeatures):
        encodec_parameters = {
            "feature_extractor.encodec." + key: value
            for key, value in model.feature_extractor.encodec.state_dict().items()
        }
        state_dict.update(encodec_parameters)
    model.load_state_dict(state_dict)
    model.eval()
    model.to(device)
    return model

However, this line is hard-coded to load pretrained encodec, preventing offline usage:

vocos/vocos/feature_extractors.py

Line 68 in c859e3b

self.encodec = encodec(pretrained=True)