Histogram of the mean specific intensity deviation in the core component for a deep learning model trained with noiseless (orange), one trained with noisy input data (blue) and clean images generated with wsclean (green). The results are near the optimal value of zero on a similar level, although there are some outliers, which is represented by the relatively high standard deviation. For wsclean the peak is shifted slightly into the negative values, which indicates a slight underestimation of the flux densities.