Source: keras_text/models/token_model.py#L0

TokenModelFactory

TokenModelFactory.__init__

__init__(self, num_classes, token_index, max_tokens, embedding_type="glove.6B.100d", \
    embedding_dims=100)

Creates a TokenModelFactory instance for building various models that operate over (samples, max_tokens) input. The token can be character, word or any other elementary token.

Args:

  • num_classes: The number of output classes.
  • token_index: The dictionary of token and its corresponding integer index value.
  • max_tokens: The max number of tokens across all documents. Depending on SequenceEncoderBase.requires_padding, this value is either used or discarded.
  • embedding_type: The embedding type to use. Set to None to use random embeddings. (Default value: 'glove.6B.100d')
  • embedding_dims: The number of embedding dims to use for representing a word. This argument will be ignored when embedding_type is set. (Default value: 100)

TokenModelFactory.build_model

build_model(self, token_encoder_model, trainable_embeddings=True, output_activation="softmax")

Builds a model using the given text_model

Args:

  • token_encoder_model: An instance of SequenceEncoderBase for encoding all the tokens within a document. This encoding is then fed into a final Dense layer for classification.
  • trainable_embeddings: Whether or not to fine tune embeddings.
  • output_activation: The output activation to use. (Default value: 'softmax')
  • Use:
  • softmax for binary or multi-class.
  • sigmoid for multi-label classification.
  • linear for regression output.

Returns:

The model output tensor.