Enable ENet with Inference Engine backend on CPU