Figure 7  Computation times across 1000 random input samples using our fully trained neural network model.