Why average pooling instead of global average pooling.

[mrgloom]

Why do you use average pooling instead of global average pooling like in paper?

layers {
  name: "CAM_pool"
  type: POOLING
  bottom: "CAM_conv"
  top: "CAM_pool"
  pooling_param {
    pool: AVE
    kernel_size: 14
    stride: 14
  }
}

[zhoubolei]

The final conv feature size is 14x14, so the AVE of 14x14 is equivalent to global average pooling. I am not sure if there is a layer in caffe called global average pooling.