Table 7 Classification accuracy, precision, recall, and F1 score of the original pre-trained model on the test set.

Models	Accuracy mean (%)	Accuracy Std	Precision	Recall	f1-score	Training time (s)	Infer time (s)
ResNet-18	82.89	1.34	0.83	0.83	0.83	3.14	0.78
ResNet-34	80.14	1.2	0.8	0.8	0.8	3.15	0.14
ResNet-50	82.91	0.79	0.83	0.83	0.83	4.53	0.23
ResNet-101	82.14	1.08	82.24	0.82	0.82	8.32	0.41
ResNet-152	81.73	0.49	81.85	0.82	0.82	12.16	0.6
MobileNet-v1	80.54	1.26	80.68	0.81	0.81	4.17	0.14
MobileNet-v2	85.58	1.1	85.55	0.86	0.86	4.72	0.22
MobileNet-v3-small	80.18	1.02	79.97	0.8	0.8	4.66	0.22
MobileNet-v3-large	84.42	0.7	84.35	0.84	0.84	5.39	0.26
MobileNet-v4	83.2	0.92	83.18	0.83	0.83	10.87	0.53
EfficientNet-B0	82.35	4.39	82.37	0.82	0.82	10.24	0.52
EfficientNet-B1	79.2	3.76	79.1	0.79	0.79	10.24	0.52
EfficientNet-B2	79.93	4.52	79.84	0.8	0.8	10.24	0.52
EfficientNet-B3	81.69	1.3	81.64	0.82	0.82	10.24	0.52
EfficientNet-B4	74.49	3.33	74.69	0.74	0.74	10.24	0.52
EfficientNet-B5	79.84	2.6	79.88	0.8	0.8	10.24	0.52
EfficientNet-B6	73.9	3.63	73.88	0.74	0.74	10.24	0.52
EfficientNet-B7	78.22	2.76	78.26	0.78	0.78	10.24	0.52
EfficientNet-B8	84.45	2.47	84.41	0.84	0.84	10.24	0.52
EfficientNet-L2	75.03	4.28	76.0	0.75	0.75	10.24	0.52
RegNetY-200MF	76.72	6.67	76.78	0.77	0.77	6.3	0.32
RegNetY-400MF	73.54	5.24	73.84	0.74	0.73	6.3	0.32
RegNetY-600MF	80.95	3.38	80.92	0.81	0.81	6.3	0.32
RegNetY-800MF	83.52	3.15	83.57	0.84	0.83	6.3	0.32
RegNetY-1.6GF	76.11	3.4	76.33	0.76	0.76	6.3	0.32
RegNetY-3.2GF	82.08	2.11	82.18	0.82	0.82	6.3	0.32
RegNetY-4.0GF	75.45	3.14	75.64	0.75	0.75	6.3	0.32
RegNetY-6.4GF	79.9	2.72	80.08	0.8	0.8	6.3	0.32
RegNetY-8.0GF	80.08	2.0	80.47	0.8	0.8	6.3	0.32
RegNetY-12GF	86.22	1.51	86.29	0.86	0.86	6.3	0.32
RegNetY-16GF	86.26	1.4	86.34	0.86	0.86	6.3	0.32
RegNetY-32GF	82.96	2.4	83.2	0.83	0.86	6.3	0.32

Quick links

Search