5 Matching Annotations
  1. Aug 2022
  2. Aug 2017
    1. First over the original image and second over the image but with the image shifted spatially by 16 pixels along both width and height

      I am confused, it means shifting by 16 pixels along width and height respectively and shifting both width and height by 16 pixel. Then the output will be [6x6 + 5x6 + 5x6 + 5x5, 1000]=[11x11,1000]?

    1. if every neuron in the network computes the same output, then they will also all compute the same gradients during backpropagation and undergo the exact same parameter updates.

      except for the weights of the first layer? As the inputs could be different.