Concatenate SSCB results from image layer with numeric CB feature data in deep learning model.

I'm using trainNetwork() for transfer learning. I've set up a concatenation layer for combining metadata as an auxilary input with our image data to help with classifications.

I'm getting the following error.

Error using trainNetwork (line 184)

Invalid network.

Caused by:

Layer 'concat': Input size mismatch. Size of input to this layer is different from the expected input size.

Inputs to this layer:

from layer 'prob' (size 1(S) × 1(S) × 8(C) × 1(B))

from layer 'features' (size 10(C) × 1(B))

How can I address this?

Answer 1

[yanqi liu]

yes，sir，is the classes number 10？may be use

layersTransfer = net.Layers(1:end-3);
numClasses = 10;
layers = [
    layersTransfer
    fullyConnectedLayer(numClasses,'WeightLearnRateFactor',20,'BiasLearnRateFactor',20)
    softmaxLayer
    classificationLayer];