def create_word_rnn_model(self, emb_dim, emb_path, vocab_word,
vocab_word_size, word_maxlen):
from keras.layers import SimpleRNN
logger.info('Building word SimpleRNN model')
input_word = Input(shape=(word_maxlen, ), name='input_word')
word_emb = Embedding(
vocab_word_size, emb_dim, mask_zero=True,
name='word_emb')(input_word)
rnn = SimpleRNN(
300,
return_sequences=True,
dropout=self.dropout,
recurrent_dropout=self.recurrent_dropout)(word_emb)
dropped = Dropout(0.5)(rnn)
mot = MeanOverTime(mask_zero=True)(dropped)
densed = Dense(self.num_outputs, name='dense')(mot)
output = Activation('sigmoid')(densed)
model = Model(inputs=input_word, outputs=output)
model.get_layer('dense').bias.set_value(self.bias)
if emb_path:
from emb_reader import EmbReader as EmbReader
logger.info('Initializing lookup table')
emb_reader = EmbReader(emb_path, emb_dim=emb_dim)
model.get_layer('word_emb').embeddings.set_value(
emb_reader.get_emb_matrix_given_vocab(
vocab_word,
model.get_layer('word_emb').embeddings.get_value()))
logger.info(' Done')
return model
def create_word_cnn_model(self, emb_dim, emb_path, vocab_word,
vocab_word_size, word_maxlen):
from aes.layers import Conv1DMask
logger.info('Building word CNN model')
input_word = Input(shape=(word_maxlen, ), name='input_word')
word_emb = Embedding(
vocab_word_size, emb_dim, mask_zero=True,
name='word_emb')(input_word)
cnn = Conv1DMask(
filters=emb_dim, kernel_size=3, padding='same')(word_emb)
dropped = Dropout(0.5)(cnn)
mot = MeanOverTime(mask_zero=True)(dropped)
densed = Dense(self.num_outputs, name='dense')(mot)
output = Activation('sigmoid')(densed)
model = Model(inputs=input_word, outputs=output)
model.get_layer('dense').bias.set_value(self.bias)
if emb_path:
from emb_reader import EmbReader as EmbReader
logger.info('Initializing lookup table')
emb_reader = EmbReader(emb_path, emb_dim=emb_dim)
model.get_layer('word_emb').embeddings.set_value(
emb_reader.get_emb_matrix_given_vocab(
vocab_word,
model.get_layer('word_emb').embeddings.get_value()))
logger.info(' Done')
return model
def create_char_cnn_model(self, emb_dim, word_maxlen, vocab_char_size,
char_maxlen):
from aes.layers import Conv1DMask
logger.info('Building character CNN model')
input_char = Input(shape=(char_maxlen, ), name='input_char')
char_emb = Embedding(
vocab_char_size, emb_dim, mask_zero=True)(input_char)
cnn = Conv1DMask(
filters=emb_dim, kernel_size=3, padding='same')(char_emb)
dropped = Dropout(0.5)(cnn)
mot = MeanOverTime(mask_zero=True)(dropped)
densed = Dense(self.num_outputs, name='dense')(mot)
output = Activation('sigmoid')(densed)
model = Model(inputs=input_char, outputs=output)
model.get_layer('dense').bias.set_value(self.bias)
logger.info(' Done')
return model
def create_char_rnn_model(self, emb_dim, word_maxlen, vocab_char_size,
char_maxlen):
from keras.layers import SimpleRNN
logger.info('Building character RNN model')
input_char = Input(shape=(char_maxlen, ), name='input_char')
char_emb = Embedding(
vocab_char_size, emb_dim, mask_zero=True)(input_char)
rnn = SimpleRNN(
300,
return_sequences=True,
dropout=self.dropout,
recurrent_dropout=self.recurrent_dropout)(char_emb)
dropped = Dropout(0.5)(rnn)
mot = MeanOverTime(mask_zero=True)(dropped)
densed = Dense(self.num_outputs, name='dense')(mot)
output = Activation('sigmoid')(densed)
model = Model(inputs=input_char, outputs=output)
model.get_layer('dense').bias.set_value(self.bias)
logger.info(' Done')
return model
def create_gate_positional_model(self, char_cnn_kernel, cnn_kernel,
emb_dim, emb_path, vocab_word,
vocab_word_size, word_maxlen,
vocab_char_size, char_maxlen):
from aes.layers import Conv1DMask, GatePositional, MaxPooling1DMask
logger.info('Building gate positional model')
input_char = Input(shape=(char_maxlen, ), name='input_char')
char_emb = Embedding(
vocab_char_size, emb_dim, mask_zero=True)(input_char)
char_cnn = Conv1DMask(
filters=emb_dim,
kernel_size=3,
padding='same')(char_emb)
char_input = MaxPooling1DMask(
pool_size=char_maxlen / word_maxlen, padding='same')(char_cnn)
input_word = Input(shape=(word_maxlen, ), name='input_word')
word_input = Embedding(
vocab_word_size, emb_dim, mask_zero=True,
name='word_emb')(input_word)
gate = GatePositional()([char_input, word_input])
final_input = Dense(50)(gate)
cnn = Conv1DMask(
filters=emb_dim,
kernel_size=3,
padding='same')(final_input)
dropped = Dropout(0.5)(cnn)
mot = MeanOverTime(mask_zero=True)(dropped)
densed = Dense(self.num_outputs, name='dense')(mot)
output = Activation('sigmoid')(densed)
model = Model(inputs=[input_char, input_word], outputs=output)
model.get_layer('dense').bias.set_value(self.bias)
if emb_path:
from emb_reader import EmbReader as EmbReader
logger.info('Initializing lookup table')
emb_reader = EmbReader(emb_path, emb_dim=emb_dim)
model.get_layer('word_emb').embeddings.set_value(
emb_reader.get_emb_matrix_given_vocab(
vocab_word,
model.get_layer('word_emb').embeddings.get_value()))
logger.info(' Done')
return model