write a Jupyter notebook a$1.$ipynb that trains and evaluates a deep learning classifier on the MNIST dataset. The notebook must use your function build$\_$deep$\_$nn to build the deep learning classifier, and then use keras$\_$tuner to define the optimal parameters of the network among these:

Number of hidden layers: $1$ to $3.$

Size of the hidden layers. To simplify, assign the same size to all hidden layers.

Dropout rate of the final hidden layer. To simplify, set a dropout rate of zero for all hidden layers except the last one.