|
Listing 1 TensorFlow Keras script for RNN implementation |
-
1:
verbose = 0 ▹ data presentation off
-
2:
epochs = 40 ▹ maximal number of epochs
-
3:
batch_size = 64 ▹ number of samples that will be propagated through the network
-
4:
n_timesteps, n_features, n_outputs = trainX.shape[1], trainX.shape[2], trainy.shape[1]
-
5:
model = Sequential()
-
6:
model.add(LSTM(60, input_shape=(n_timesteps,n_features)))
-
7:
model.add(Dropout(0.5))
-
8:
model.add(Dense(60, activation=’relu’))
-
9:
model.add(Dense(n_outputs, activation=’softmax’))
-
10:
model.compile(loss=’categorical_crossentropy’,
optimizer=’adam’,
metrics=[’accuracy’])
|