mindmeld.models.taggers.embeddings module

class mindmeld.models.taggers.embeddings.CharacterSequenceEmbedding(sequence_padding_length, token_embedding_dimension=None, max_char_per_word=None)

Bases: object

CharacterSequenceEmbedding encodes a sequence of words into a sequence of fixed dimension real-numbered vectors by mapping each character in the words as vectors.

encode_sequence_of_tokens(token_sequence)

Encodes a sequence of tokens into real value vectors.

Parameters:token_sequence (list) -- A sequence of tokens. Returns:Encoded sequence of tokens. Return type:(list)

save_embeddings()

Save extracted embeddings to historic pickle file.

class mindmeld.models.taggers.embeddings.WordSequenceEmbedding(sequence_padding_length, token_embedding_dimension=None, token_pretrained_embedding_filepath=None, use_padding=True)

Bases: object

WordSequenceEmbedding encodes a sequence of words into a sequence of fixed dimension real-numbered vectors by mapping each word as a vector.

encode_sequence_of_tokens(token_sequence)

Encodes a sequence of tokens into real value vectors.

Parameters:token_sequence (list) -- A sequence of tokens. Returns:Encoded sequence of tokens. Return type:(list)

save_embeddings()

Save extracted embeddings to historic pickle file.