Applied-Deep-Learning-with-TensorFlow

Applied-Deep-Learning-with-TensorFlow copied to clipboard

Hey! First of all thanks for the tutorial, I have a problem and not able to resolve it.

The accuracy what I have got is around 63% it's not good enough has mentioned in the video so I thought of adding more bidirectional LSTM layers to the existing network to check out the result but I am getting this error. Please help me to resolve this issue.

sequence = Input(shape = (max_len, ), dtype = np.int32)
embedding = Embedding(max_features, embedding_dim, input_length = max_len)(sequence)
batch_norm = BatchNormalization()(embedding)

bi_lstm = Bidirectional(LSTM(lstm_layer_dim))(batch_norm)
drop_out = Dropout(0.25)(bi_lstm)
bi_lstm_1 = Bidirectional(LSTM(lstm_layer_dim_1))(drop_out)
drop_out_1 = Dropout(0.25)(bi_lstm_1)
preds = Dense(2, activation = 'softmax')(drop_out_1)

model = Model(sequence, preds)
model.summary()

The Error which I am getting is this,

---

IndexError Traceback (most recent call last) in () 6 bi_lstm = Bidirectional(LSTM(lstm_layer_dim))(batch_norm) 7 drop_out = Dropout(0.25)(bi_lstm) ----> 8 bi_lstm_1 = Bidirectional(LSTM(lstm_layer_dim_1))(drop_out) 9 drop_out_1 = Dropout(0.25)(bi_lstm_1) 10 preds = Dense(2, activation = 'softmax')(drop_out_1)

C:\Users\Akshath\AppData\Local\Continuum\Anaconda3\envs\py35\lib\site-packages\keras\engine\topology.py in call(self, inputs, **kwargs) 573 'layer.build(batch_input_shape)') 574 if len(input_shapes) == 1: --> 575 self.build(input_shapes[0]) 576 else: 577 self.build(input_shapes)

C:\Users\Akshath\AppData\Local\Continuum\Anaconda3\envs\py35\lib\site-packages\keras\layers\wrappers.py in build(self, input_shape) 322 def build(self, input_shape): 323 with K.name_scope(self.forward_layer.name): --> 324 self.forward_layer.build(input_shape) 325 with K.name_scope(self.backward_layer.name): 326 self.backward_layer.build(input_shape)

C:\Users\Akshath\AppData\Local\Continuum\Anaconda3\envs\py35\lib\site-packages\keras\layers\recurrent.py in build(self, input_shape) 1015 1016 batch_size = input_shape[0] if self.stateful else None -> 1017 self.input_dim = input_shape[2] 1018 self.input_spec[0] = InputSpec(shape=(batch_size, None, self.input_dim)) 1019

IndexError: tuple index out of range