Suppose we a convolutional neural network.Inputs to convolutions are properly padded so that convolutions do not reduce the size. The kernel dimension (depth) matches the size of the previous layers output. Bias term is added convolution outputs.

X = 7

Y = 0

Suppose we a convolutional neural network. Inputs to convolutions are properly padded so that convolutions do not reduce the size. The kernel dimension (depth) matches the size of the previous layers output. Bias term is added convolution outputs. The layers and descriptions are given below. Calculate the number of parameters at each layer. Number of parameters Layer Layer 1 (input layer) Layer 2 Layer 3 Layer 4 Description 3 channel image of size 100x100 Convolutional layer with 10 filters. Kernel size is (X+1)x(X+1). Activation function is ReLU MaxPool layer with stride 2. Convolutional layer with 5 filters. Kernel size is (Y+1)x(Y+1). Activation function is ReLU. Flatten Fully connected layer with 10 units. Bias term is added. Fully connected layer with 2 units. Bias term is added Softmax Layer 5 Layer 6 Layer 7 Layer 8 (output layer)