Skip to content

bimpm

allennlp_models.pair_classification.models.bimpm

[SOURCE]

BiMPM (Bilateral Multi-Perspective Matching) model implementation.

BiMpm#

@Model.register("bimpm")
class BiMpm(Model):
 | def __init__(
 |     self,
 |     vocab: Vocabulary,
 |     text_field_embedder: TextFieldEmbedder,
 |     matcher_word: BiMpmMatching,
 |     encoder1: Seq2SeqEncoder,
 |     matcher_forward1: BiMpmMatching,
 |     matcher_backward1: BiMpmMatching,
 |     encoder2: Seq2SeqEncoder,
 |     matcher_forward2: BiMpmMatching,
 |     matcher_backward2: BiMpmMatching,
 |     aggregator: Seq2VecEncoder,
 |     classifier_feedforward: FeedForward,
 |     dropout: float = 0.1,
 |     initializer: InitializerApplicator = InitializerApplicator(),
 |     **kwargs
 | ) -> None

This Model implements BiMPM model described in Bilateral Multi-Perspective Matching for Natural Language Sentences by Zhiguo Wang et al., 2017. Also please refer to the TensorFlow implementation and PyTorch implementation.

Registered as a Model with name "bimpm".

Parameters

  • vocab : Vocabulary
  • text_field_embedder : TextFieldEmbedder
    Used to embed the premise and hypothesis TextFields we get as input to the model.
  • matcher_word : BiMpmMatching
    BiMPM matching on the output of word embeddings of premise and hypothesis.
  • encoder1 : Seq2SeqEncoder
    First encoder layer for the premise and hypothesis
  • matcher_forward1 : BiMPMMatching
    BiMPM matching for the forward output of first encoder layer
  • matcher_backward1 : BiMPMMatching
    BiMPM matching for the backward output of first encoder layer
  • encoder2 : Seq2SeqEncoder
    Second encoder layer for the premise and hypothesis
  • matcher_forward2 : BiMPMMatching
    BiMPM matching for the forward output of second encoder layer
  • matcher_backward2 : BiMPMMatching
    BiMPM matching for the backward output of second encoder layer
  • aggregator : Seq2VecEncoder
    Aggregator of all BiMPM matching vectors
  • classifier_feedforward : FeedForward
    Fully connected layers for classification.
  • dropout : float, optional (default = 0.1)
    Dropout percentage to use.
  • initializer : InitializerApplicator, optional (default = InitializerApplicator())
    If provided, will be used to initialize the model parameters.

forward#

class BiMpm(Model):
 | ...
 | @overrides
 | def forward(
 |     self,
 |     premise: TextFieldTensors,
 |     hypothesis: TextFieldTensors,
 |     label: torch.LongTensor = None,
 |     metadata: List[Dict[str, Any]] = None
 | ) -> Dict[str, torch.Tensor]

Parameters

  • premise : TextFieldTensors
    The premise from a TextField
  • hypothesis : TextFieldTensors
    The hypothesis from a TextField
  • label : torch.LongTensor, optional (default = None)
    The label for the pair of the premise and the hypothesis
  • metadata : List[Dict[str, Any]], optional (default = None)
    Additional information about the pair

Returns

  • An output dictionary consisting of:

  • logits : torch.FloatTensor
    A tensor of shape (batch_size, num_labels) representing unnormalised log probabilities of the entailment label.

  • loss : torch.FloatTensor, optional
    A scalar loss to be optimised.

make_output_human_readable#

class BiMpm(Model):
 | ...
 | @overrides
 | def make_output_human_readable(
 |     self,
 |     output_dict: Dict[str, torch.Tensor]
 | ) -> Dict[str, torch.Tensor]

Converts indices to string labels, and adds a "label" key to the result.

get_metrics#

class BiMpm(Model):
 | ...
 | @overrides
 | def get_metrics(self, reset: bool = False) -> Dict[str, float]

default_predictor#

class BiMpm(Model):
 | ...
 | default_predictor = "textual_entailment"