How to concat the output of a pooling layer and inner-product layer in CAFFE?

[fby...@gmail.com]

While trying to concat the output of a pooling layer (pooling a convolution layer) with the outputs of an inner product layer, I got the following error:

   

 Check failed: num_axes == bottom[i]->num_axes() (4 vs. 2) All inputs must have the same #axes.

I would like to concat both and feed them into another (fully-connected) inner-product layer.

I don't mind flattening the output of the conv layer.

Can anyone suggests how to solve this?

my concat layer definition is:

    layer {
      type: "Concat"
      bottom: "pool3"
      bottom: "ip1_prior"
      top: "ip1_combine"
      name: "concat"
    }

Thank you!

[fby...@gmail.com]

Solved my issue by flattening the conv. layer output.

e.g. by adding a Flatten layer between the pool and concat layers:

layer {
  type: "Flatten"
  bottom: "pool3"
  top: "pool3flat"
  name: "p3flat"
}


layer {

  type: "Concat"
  bottom: "pool3flat"