gm.text.Gemma2Tokenizer

gm.text.Gemma2Tokenizer#

class gemma.gm.text.Gemma2Tokenizer(
path: str | os.PathLike = 'gs://gemma-data/tokenizers/tokenizer_gemma2.model',
*,
custom_tokens: dict[int,
str] = <factory>,
)[source]

Bases: gemma.gm.text._tokenizer.Tokenizer

Gemma 2 的分词器。

path: str | os.PathLike = 'gs://gemma-data/tokenizers/tokenizer_gemma2.model'
special_tokens[source]

别名: gemma.gm.text._tokenizer._Gemma2SpecialTokens

VERSION: ClassVar[int] = 2