표 1. | Table 1. 모델별 성능 비교 | Comparative analysis of model performances.

Model	Inference time (sec)	GPU Memory (MByte)	Top-1 Acc (%)	Top-2 Acc (%)	Top-5 Acc (%)
ShuffleNet V2 x0.5	2.99	592	81.83	93.18	98.46
ShuffleNet V2 x2.0	3.75	602	87.38	95.68	99.10
ResNet-50	5.54	1,102	89.89	96.25	99.14
ResNet-152	8.55	1,420	89.84	96.66	99.25
ResNext-101-32x8d	11.10	1,626	89.98	96.49	99.30
MobileNet V3 Small	3.75	668	88.11	95.56	98.85
MobileNet V3 Large	4.31	925	89.36	95.85	98.92
Wide ResNet-50-2	6.56	1,363	89.97	96.47	99.20
Wide ResNet-101-2	10.93	1,709	89.73	96.46	99.34
Densenet-169	6.63	971	90.24	96.56	99.24
Densenet-201	7.84	984	89.90	96.61	99.41