I am trying to implement the behavior of a ResNet (Residual Neural Network) in tensorflow.js. I am hoping someone who knows what they are doing can point me in the right direction. Would the following code effectively make layers 3,4,5, and into a residual block? More specifically, does the concatenate layer I have included function like the identity skip connection described in the literature? Does TFJS automatically know how to pass the backwards propagation signal through the concat layer?