Write a function that calculates the output of a toy 2-layers NN (foward) to perform a binary classification task.

Assume the following:

*The NN is defined by one hidden layer with 4 units and the output layer.

*The function takes an input vector (1XM), the corresponding weight matrices and the bias vector for each layer.

*Choose an activation function for the hidden layer. You don't need to implement the activation functions. You can assume you can call the function or use the corresponding library.

*The function returns the output of the last layer.