Fig. 16From: 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 LSTM modelBack to article page