Fig. 1From: Deep convolutional architectures for extrapolative forecasts in time-dependent flow problemsTraining and validation methodBack to article page