Answer the question carefully...

Consider a convolutional network with the following configuration. Calculate (show the calculation) the number of trainable parameters per layer and in total. You should assume that none of the layers use parameters for bias. Details Layer Type # Number of parameters 1 Input Layer 50 X 50 CMYK Image N Normalisation layer Input - it's mean (channel-wise) 3 3 Convolutional 6, 3x3 filters 4 Activation Relu activation 01 5 Convolutional 14, 1x1 filters 0 Activation Rell activation 8 00 Flattening 9 Fully connected 37 neurons, Activation - Sigmoid - Total number of parameters