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

Table 4

Executing time of conflicting situations on learning rate.

Methods	Decay	1R1D	2R2D	3R3D	5R5D
RMSprop	64	69	70	71	72
Adam	70	78	79	81	83
NAdam [29]	79	91	92	93	95
AdaMax [5]	74	89	90	92	94
Adadelta	83	104	106	108	112
Adagrad	82	107	108	109	111

Unit: ms/batch; kRkD: raise k times and decrease k times on learning rate for each iteration.