def __init__(self):
super().__init__(
inputter=inputters.MixedInputter(
[
inputters.WordEmbedder(embedding_size=100),
inputters.CharConvEmbedder(
embedding_size=30,
num_outputs=30,
kernel_size=3,
stride=1,
dropout=0.5,
),
],
dropout=0.5,
),
encoder=encoders.RNNEncoder(
num_layers=1,
num_units=400,
bidirectional=True,
dropout=0.5,
residual_connections=False,
cell_class=tf.keras.layers.LSTMCell,
),
crf_decoding=True,
)
def testBidirectionalRNNEncoder(self, cell_class):
encoder = encoders.RNNEncoder(3, 20, bidirectional=True, cell_class=cell_class)
inputs = tf.random.uniform([4, 5, 10])
lengths = tf.constant([4, 3, 5, 2])
outputs, states, _ = encoder(inputs, sequence_length=lengths, training=True)
self.assertListEqual(outputs.shape.as_list(), [4, 5, 40])
self.assertEqual(len(states), 3)
def __init__(self):
super(LuongAttention, self).__init__(
source_inputter=inputters.WordEmbedder(embedding_size=512),
target_inputter=inputters.WordEmbedder(embedding_size=512),
encoder=encoders.RNNEncoder(num_layers=4,
num_units=1000,
dropout=0.2,
residual_connections=False,
cell_class=tf.keras.layers.LSTMCell),
decoder=decoders.AttentionalRNNDecoder(
num_layers=4,
num_units=1000,
bridge_class=layers.CopyBridge,
attention_mechanism_class=tfa.seq2seq.LuongAttention,
cell_class=tf.keras.layers.LSTMCell,
dropout=0.2,
residual_connections=False))
def __init__(self):
# pylint: disable=bad-continuation
super(LstmCnnCrfTagger, self).__init__(
inputter=inputters.MixedInputter([
inputters.WordEmbedder(embedding_size=100),
inputters.CharConvEmbedder(embedding_size=30,
num_outputs=30,
kernel_size=3,
stride=1,
dropout=0.5)
],
dropout=0.5),
encoder=encoders.RNNEncoder(num_layers=1,
num_units=400,
bidirectional=True,
dropout=0.5,
residual_connections=False,
cell_class=tf.keras.layers.LSTMCell),
crf_decoding=True)
def __init__(self):
super(NMTMediumV1, self).__init__(
source_inputter=inputters.WordEmbedder(embedding_size=512),
target_inputter=inputters.WordEmbedder(embedding_size=512),
encoder=encoders.RNNEncoder(num_layers=4,
num_units=256,
bidirectional=True,
residual_connections=False,
dropout=0.3,
reducer=layers.ConcatReducer(),
cell_class=tf.keras.layers.LSTMCell),
decoder=decoders.AttentionalRNNDecoder(
num_layers=4,
num_units=512,
bridge_class=layers.CopyBridge,
attention_mechanism_class=tfa.seq2seq.LuongAttention,
attention_layer_activation=None,
cell_class=tf.keras.layers.LSTMCell,
dropout=0.3,
residual_connections=False))
