Photometric parameters derived in this work compared against test datasets. Top row: 2D histograms showing the performance of the trained photometric parameter inference network on all 10 000 clusters from the validation dataset. The mean output uncertainty is shown with white error bars. As indicated by the dashed lines, predicted values on the y axis should be equal to true values on the x axis. The root mean square error (RMSE) and mean absolute error in terms of output network uncertainty (MAE) are given in the top left. All plots and the RMSE are in units of magnitude other than on age plots which are logarithms of cluster age in years. Other rows: comparison between network predicted parameters and ages, extinctions, and distance moduli for 247, 1753, and 1206 clusters in common with the catalogues of Bossini et al. (2019), Cantat-Gaudin et al. (2020), and Kharchenko et al. (2013) respectively. Points are shaded based on the differential extinction we infer for each cluster.