In your implementation of bilinear VGG in caffe, I found out that num_output for bilinear_layer is at 8192. But I do not understand how is it so? Because no of output for this layer should be equal to 512x512 if we take outer product of the preceding layer output with itself.
I now understand this. You have used compact bilinear pooling also. If possible can you also pls share the link on implementation of "bilinear layers" in caffe.
In your implementation of bilinear VGG in caffe, I found out that
num_output
forbilinear_layer
is at8192
. But I do not understand how is it so? Because no of output for this layer should be equal to512x512
if we take outer product of the preceding layer output with itself.