SnliReader( self, tokenizer: allennlp.data.tokenizers.tokenizer.Tokenizer = None, token_indexers: Dict[str, allennlp.data.token_indexers.token_indexer.TokenIndexer] = None, combine_input_fields: bool = None, kwargs, ) -> None
Reads a file from the Stanford Natural Language Inference (SNLI) dataset. This data is formatted as jsonl, one json-formatted instance per line. The keys in the data are "gold_label", "sentence1", and "sentence2". We convert these keys into fields named "label", "premise" and "hypothesis", along with a metadata field containing the tokenized strings of the premise and hypothesis.
Registered as a
DatasetReader with name "snli".
- tokenizer :
Tokenizer, optional (default=
- We use this
Tokenizerfor both the premise and the hypothesis. See :class:
- token_indexers :
Dict[str, TokenIndexer], optional (default=
- We similarly use this for both the premise and the hypothesis. See :class:
- combine_input_fields :
bool, optional (default=
isinstance(tokenizer, PretrainedTransformerTokenizer)) If False, represent the premise and the hypothesis as separate fields in the instance. If True, tokenize them together using
tokenizer.tokenize_sentence_pair()and provide a single
tokensfield in the instance.
SnliReader.text_to_instance( self, premise: str, hypothesis: str, label: str = None, ) -> allennlp.data.instance.Instance
Does whatever tokenization or processing is necessary to go from textual input to an
Instance. The primary intended use for this is with a
~allennlp.predictors.predictor.Predictor, which gets text input as a JSON
object and needs to process it to be input to a model.
The intent here is to share code between :func:
_read and what happens at
model serving time, or any other time you want to make a prediction from new data. We need
to process the data in the same way it was done at training time. Allowing the
DatasetReader to process new text lets us accomplish this, as we can just call
DatasetReader.text_to_instance when serving predictions.
The input type here is rather vaguely specified, unfortunately. The
have to make some assumptions about the kind of
DatasetReader that it's using, in order
to pass it the right information.