Skip to content

whitespace_tokenizer

allennlp.data.tokenizers.whitespace_tokenizer

[SOURCE]


WhitespaceTokenizer

@Tokenizer.register("whitespace")
@Tokenizer.register("just_spaces")
class WhitespaceTokenizer(Tokenizer)

A Tokenizer that assumes you've already done your own tokenization somehow and have separated the tokens by spaces. We just split the input string on whitespace and return the resulting list.

Note that we use text.split(), which means that the amount of whitespace between the tokens does not matter. This will never result in spaces being included as tokens.

Registered as a Tokenizer with name "whitespace" and "just_spaces".

tokenize

class WhitespaceTokenizer(Tokenizer):
 | ...
 | def tokenize(self, text: str) -> List[Token]