Classification should be a VisionLabeledModule #965
Description
Currently the classification doesn't allow to pass list of labels per model. It should similarly to object detection and segmentation extend the VisionLabeledModule and allow for such configuration.