Table 3. Optimized hyperparameters and results referring to the N_1b network, obtained from a grid scan.

Parameter	Value
Width	2
Depth	230
Batch size	512
Regularization parameter (l2)	1 × 10−8
Learning rate	5 × 10−5
Dropout function parameter	0.1
Final loss	3.4 × 10−3
Final validation loss	2.9 × 10−3