vllm_omni.model_executor.models.indextts2.utils.front ¶

convert_ids_to_tokens(ids: list[int] | int)

convert_tokens_to_ids(tokens: list[str] | str) -> list[int]

decode(ids: list[int] | int, do_lower_case=False, **kwargs)

encode(text: str, **kwargs)

get_vocab()

split_segments(
    tokenized: list[str],
    max_text_tokens_per_segment=120,
    quick_streaming_tokens=0,
) -> list[list[str]]

split_segments_by_token(
    tokenized_str: list[str],
    split_tokens: list[str],
    max_text_tokens_per_segment: int,
    quick_streaming_tokens: int = 0,
) -> list[list[str]]

tokenize(text: str) -> list[str]