Figure 2From: Operator compression with deep neural networksDevelopment of loss functional \(\mathcal{J}\) during 20 epochs of the network trainingBack to article page