Question $6[2$ pts$]$ The Dense Layer Classification $[$Notebook$,$ html$]$ in the Canvas shows

detailed procedures about how to use Oivetti face dataset $($from AT&T$)$ to train Neural Network

classifiers for face classification. The dataset in the Canvas also provides $$olivetti$\_$faces.npy$$

and $$olivetti$\_$faces$\_$target.npy$,$ which includes $400$ faces in $40$ classes $(40$ different person$).$

Please implement following face recognition task using Neural Networks.

$$ Please show at least one face images for each class in the Oivetti face dataset$100.$

Randomly split the dataset into $60\%$ training and $40\%$ test samples. Train a one$-$hidden

layer neural network with $10$ hidden nodes. Report the classification accuracy of the

classifier on the test set $[1$ pt$]$

$$ Please use one time $10-$fold cross validation to compare the performance of different

neural network architectures, including $(1)$ one$-$hidden layer NN with $10$ hidden nodes,

$(2)$ one$-$hidden layer NN with $50$ hidden nodes, $(3)$ one$-$hidden layer NN with $500$ hidden

nodes, and $(4)$ two$-$hidden layer NN with $50$ hidden nodes $(1$st layer$)$ and $10$ hidden nodes

$(2$n layer$).$ Please report and compare the cross$-$validation accuracy of the four neural

networks, and conclude which classifier has the best performance $[1$ pt$].$