Table 5.

Average accuracy for the second version model using the COV-PEN dataset with the first 50 layers frozen, epochs = 15, optimizer = sgmd, and batch size = 128.

Learning Rate	Ensemble Using Several Runs
	Run 1	Run 2	Run 3
0.0002	0.933691756	0.937275986	0.933691756
0.0004	0.939068100	0.935483871	0.903942652
0.0006	0.942652329	0.917562724	0.907526882
0.0008	0.892473118	0.716845878	0.808960573
0.001	0.747311828	0.921146953	0.808960573