We shall now build the classification network, with a hidden layer(with 30 neurons and selu activation function) and an output layer( with 1 neuron since we have to predict only one value).
We shall also specify the loss function we have defined - the huber_fn and the optimizer to use, while compiling.
Note:
kernel_initializer defines the way to set the initial random weights of Keras layers.Set "selu" as the activation and "lecun_normal" as the kernel_initializer for the following model.
model = keras.models.Sequential([
keras.layers.Dense(30, activation=<< your code comes here >>,
kernel_initializer=<< your code comes here >>,
input_shape=X_train.shape[1:]),
keras.layers.Dense(1),
])
Compile the model using compile.
Set loss to be huber_fn which is the custom loss function we defined
We will be using "nadam" optimizer and "mae" metric.
model.<< your code comes here >>(loss=<< your code comes here >>, optimizer="nadam", metrics=["mae"])
Note - Having trouble with the assessment engine? Follow the steps listed here
Loading comments...