$($CNN$)$ I always remember to Thumbs Up correct posts!

$-$First, compute the dimension of each feature maps resulted from

convolution$/$max$-$pooling, and neuron number of linear layers $($LL$)$;

$-$Second, compute the number of parameters of the neural network $($no padding allowed$)(10$

points$)$;

$-$Third, implement the architecture using PyTorch