[Retracted] Gradient Descent Optimization in Deep Learning Model Training Based on Multistage and Method Combination Strategy

Figure 3

Combination strategy. Each method has its own color to mark itself. (a) Inner combination case; (b) mixed combination case.

(a)

(b)