Use ffmpeg read data (#1349)

夜雨飘零

2024-02-02 c2f174cd7811a7a11e6227ecb371887f97fd66d3

 funasr/tokenizer/phoneme_tokenizer.py

@@ -13,7 +13,7 @@
from funasr.tokenizer.abs_tokenizer import AbsTokenizer


g2p_choices = [
g2p_classes = [
    None,
    "g2p_en",
    "g2p_en_no_space",
@@ -363,7 +363,6 @@
        non_linguistic_symbols: Union[Path, str, Iterable[str]] = None,
        space_symbol: str = "<space>",
        remove_non_linguistic_symbols: bool = False,
        **kwargs,
    ):
        if g2p_type is None:
            self.g2p = split_by_space