The class implements a layer that finds the maximum among the elements that are at the same position in all input blobs.
This layer has no settings.
This layer has no trainable parameters.
The layer should have at least two inputs. The blobs of all inputs should have the same dimensions.
The layer has one output of the same size as each of the inputs. The output blob contains the element-wise maximum of the input blobs.