sentence_model.py 文件源码

def __init__(self, num_classes, token_index, max_sents, max_tokens,
                 embedding_type='glove.6B.100d', embedding_dims=100):
        """Creates a `SentenceModelFactory` instance for building various models that operate over
        (samples, max_sentences, max_tokens) input.

        Args:
            num_classes: The number of output classes.
            token_index: The dictionary of token and its corresponding integer index value.
            max_sents: The max number of sentences in a document.
            max_tokens: The max number of tokens in a sentence.
            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)
        """
        self.num_classes = num_classes
        self.token_index = token_index
        self.max_sents = max_sents
        self.max_tokens = max_tokens

        # This is required to make TimeDistributed(word_encoder_model) work.
        # TODO: Get rid of this restriction when https://github.com/fchollet/keras/issues/6917 resolves.
        if self.max_tokens is None:
            raise ValueError('`max_tokens` should be provided.')

        if embedding_type is not None:
            self.embeddings_index = get_embeddings_index(embedding_type)
            self.embedding_dims = self.embeddings_index.values()[0].shape[-1]
        else:
            self.embeddings_index = None
            self.embedding_dims = embedding_dims