Question No:

You come up with a CNN classifier. For eachlayer, calculate the number of weights, number biases and the size of the associated feature maps.

Marks

The notation follows the convention:

CONV denotes a convolutional layer with filters, each them of size Padding and stride parameters are always and respectively.

indicates a pooling layer with stride and padding

stands for a fullycomected layer with neurons.

tableLayertableActivation mapdimensionsNumber of weights,Number of BiasesINPUTCONVPOOLCONVPOOLCONVPOOLFC