Consider the following simple CNN architecture:

$[6$ Marks$]$

Input: $64643($RGB image$),$

Convolutional Layer $1$: $[64$ filters of size $=22,$ stride $=2,$ padding $=$ 'valid'$],$

Batch Normalization Layer: Applied after Conv Layer $1,$

Pooling Layer $1$: $[$Max pooling with filter size $=22,$ stride $=2,$ padding $=$ 'valid' $],$

Convolutional Layer $2$: $[32$ filters of size $=33,$ stride $=1,$ padding $=$ 'valid'$],$

Pooling Layer $2$: $[$Average pooling with filter size $=44,$ stride $=2,$ padding $=\text{'}$ valid'$],$

Flatten, Fully Connected Layer: $[128$ neurons$],$

Output Layer: $[10$ neurons $($for classification$)]$

For each layer of the architecture, calculate the number of the feature maps $/$ neurons, size of the feature maps, trainable parameters, and

non$-$trainable parameters?