Gradio Demo
Any chance to have a working demo up on Huggingface spaces?
Does Huggingface support fine-tuning demos (for ASR, LID, etc) ? This model by itself cannot do anything.
TypeError Traceback (most recent call last)
/tmp/ipykernel_37/3091763052.py in <cell line: 0>()
11 os.environ["CUDA_VISIBLE_DEVICES"] = "0"
12
---> 13 xeus_model, xeus_train_args = SSLTask.build_model_from_file(
14 None,
15 os.environ["MODEL_PATH"],
/usr/local/lib/python3.11/dist-packages/espnet2/tasks/abs_task.py in build_model_from_file(cls, config_file, model_file, device)
2135 args = yaml.safe_load(f)
2136 args = argparse.Namespace(**args)
-> 2137 model = cls.build_model(args)
2138 if not isinstance(model, AbsESPnetModel):
2139 raise RuntimeError(
/usr/local/lib/python3.11/dist-packages/espnet2/tasks/ssl.py in build_model(cls, args)
404 # Extract features in the model
405 frontend_class = frontend_choices.get_class(args.frontend)
--> 406 frontend = frontend_class(**args.frontend_conf)
407 input_size = frontend.output_size()
408 else:
TypeError: CNNFrontend.init() got an unexpected keyword argument 'normalize_output'