I want to know how I can calculate the output layer from (number of neurons and shape of that) of deep neural network layers. Suppose we have 32 * 32 image input of 2D convolution layer with parameters:
num_filters = 96,
kernel_size = 5,
stride = 1,
pad = 2.
How do I calculate output form or for pooling layer or other layers existing in deep neural networks?