Merge branch 'main' of https://github.com/alibaba-damo-academy/FunASR into ...

雾聪

2023-08-07 f8d1c79fe355efb18ae49e4363307dfec3ab89ce

 funasr/build_utils/build_diar_model.py

@@ -192,18 +192,22 @@

def build_diar_model(args):
    # token_list
    if isinstance(args.token_list, str):
        with open(args.token_list, encoding="utf-8") as f:
            token_list = [line.rstrip() for line in f]
    if args.token_list is not None:
        if isinstance(args.token_list, str):
            with open(args.token_list, encoding="utf-8") as f:
                token_list = [line.rstrip() for line in f]

        # Overwriting token_list to keep it as "portable".
        args.token_list = list(token_list)
    elif isinstance(args.token_list, (tuple, list)):
        token_list = list(args.token_list)
            # Overwriting token_list to keep it as "portable".
            args.token_list = list(token_list)
        elif isinstance(args.token_list, (tuple, list)):
            token_list = list(args.token_list)
        else:
            raise RuntimeError("token_list must be str or list")
        vocab_size = len(token_list)
        logging.info(f"Vocabulary size: {vocab_size}")
    else:
        raise RuntimeError("token_list must be str or list")
    vocab_size = len(token_list)
    logging.info(f"Vocabulary size: {vocab_size}")
        token_list = None
        vocab_size = None

    # frontend
    if args.input_size is None:
@@ -212,16 +216,14 @@
            frontend = frontend_class(cmvn_file=args.cmvn_file, **args.frontend_conf)
        else:
            frontend = frontend_class(**args.frontend_conf)
        input_size = frontend.output_size()
    else:
        args.frontend = None
        args.frontend_conf = {}
        frontend = None
        input_size = args.input_size

    # encoder
    encoder_class = encoder_choices.get_class(args.encoder)
    encoder = encoder_class(input_size=input_size, **args.encoder_conf)
    encoder = encoder_class(**args.encoder_conf)

    if args.model == "sond":
        # data augmentation for spectrogram
@@ -294,7 +296,7 @@
            **args.model_conf,
        )

    elif args.model_name == "eend_ola":
    elif args.model == "eend_ola":
        # encoder-decoder attractor
        encoder_decoder_attractor_class = encoder_decoder_attractor_choices.get_class(args.encoder_decoder_attractor)
        encoder_decoder_attractor = encoder_decoder_attractor_class(**args.encoder_decoder_attractor_conf)