Image Classification on TinyImageNet (Specific Robustness Metrics) benchmark leaderboard · SOTA2 Research

Image Classification on TinyImageNet (Specific Robustness Metrics)

73.74Clean AccuracyWRN50

Evaluation Results

Method	Links
WRN502025.11		73.74	59.88	93.6	96.19	93.43	4.79	3.18	0	72.31	54.85	99.76	99.65
ResNet502025.11		70.39	64.46	92.09	95.19	91.9	3.9	2.12	0	72.79	42.36	99.74	99.76
VGG162025.11		66.95	59.01	93.07	95.7	92.87	4.12	0.21	0	70.71	42.3	100.16	100.1
ResNet182025.11		56.99	53.2	92.26	95.17	92.07	0.45	0.52	0	71.61	33.78	99.96	99.33
WRN502025.11		33.37	88.95	99.84	99.93	99.81	30.33	25.77	21.22	88.97	66.5	96.19	99.55
ResNet502025.11		27.94	93.21	99.79	99.72	99.76	39.3	32.14	27.27	89.08	73.51	95.85	99.79
ResNet182025.11		21.34	91.06	99.7	99.78	99.73	44.05	32.71	27.65	90.58	72.91	94.89	99.39