Comparing activation functions

Comparing activation functions involves a bit of coding, but nothing you can't do!

You will try out different activation functions on the multi-label model you built for your irrigation machine in chapter 2. The function get_model() returns a copy of this model and applies the activation function, passed on as a parameter, to its hidden layer.

You will build a loop that goes through several activation functions, generates a new model for each and trains it. Storing the history callback in a dictionary will allow you to compare and visualize which activation function performed best in the next exercise!

X_train, X_test, y_train, y_test are ready for you when training your model.

Fill up the activation functions array with relu,leaky_relu, sigmoid, and tanh.
Get a new model for each iteration with get_model() passing the current activation function as a parameter.
Fit your model providing the train and validation_data, use 20 epochs and set verbose to 0.

Take Hint (-30 XP)

Exercise

Comparing activation functions

Instructions