Question $7[3$ pts$]$: Please follow $$Dense network for image classification $[$Notebook$,$ html$]$ to create a dense layer network image classifiers and validate its performances.

$$ Please download the cat vs$.$ dog images $($dogvscat$1000.$zip$)$ from Canvas

$$ Unzip the downloaded $($zip$)$ file. There are $1000$ dog and cat images $(500$ for each

category$)$ in the unzipped folder.

$$ Use $10-$fold cross validation to split the $1000$ images into training vs$.$ test set. Train

neural networks using training set, and report their performance on the test set.

$$ Create a feedforward neural network with one or two hidden layers $($you can determine the number of hidden nodes for each layer, but the number of hidden nodes for each layer should be at least $50).$ Train the network on the training set and report its performance on the test. Report at least one loss plot with respect to the number of epochs. Report $10-$fold cross validation accuracy, including mean and standard deviation $[1\mathrm{.}5$ pts$]$

$$ For the same network structure created above, please change the network structure by adding batch normalization and$/$or dropout layer $($at different layers$).$ Report $10-$fold cross validation performance of the network by adding batch normalization and dropout layer, respectively $($you can decide which layer$($s$)$ to add batch normalization and$/$or dropout$).$ Conclude batch normalization and drop$-$out impact on the neural network performance.