np_train_f_t = np.append(np_train_f_t, np_t) if np_train_f_v is None: np_train_f_v = np_v else: np_train_f_v = np.append(np_train_f_v, np_v) if np_train_f_ic is None: np_train_f_ic = np_ic else: np_train_f_ic = np.append(np_train_f_ic, np_ic) # Normalizers t_normalizer = Normalizer() v_normalizer = Normalizer() i_normalizer = Normalizer() t_normalizer.parametrize(np_t) v_normalizer.parametrize(np.array(train_vs)) i_normalizer.parametrize(np.array(train_ics)) # Train data normalization np_norm_train_u_t = t_normalizer.normalize(np_train_u_t) np_norm_train_u_v = v_normalizer.normalize(np_train_u_v) np_norm_train_u_ic = i_normalizer.normalize(np_train_u_ic) np_norm_train_f_t = t_normalizer.normalize(np_train_f_t) np_norm_train_f_v = v_normalizer.normalize(np_train_f_v) np_norm_train_f_ic = i_normalizer.normalize(np_train_f_ic) # PINN instancing hidden_layers = [9, 9] learning_rate = 0.001