Fig. 20From: Deep convolutional architectures for extrapolative forecasts in time-dependent flow problemsStoker’s problem, \(L_2\) relative error of the auto-regressive predictions with increasing time for the CNN modelBack to article page