MergeVertex behaviour with both Convolutional and Regular layers

DL4J
1

I’ve read the MergeVertex concatenates the activations, and that the concaenation is along the depth in case of Convolutional networks. I’d like to know if this last scenario is the case also if we are merging a ConvolutionLayer with a DenseLayer, thank you very much.

2

@EquanimeAugello you’ll need to reshape everything to match then pass it to merge.