Classification on CIFAR100C (test) benchmark leaderboard · SOTA2 Research

Classification on CIFAR100C (test)

34.76Top-1 AccuracyAdaFocal

Evaluation Results

Method	Links
AdaFocal2026.05		34.76	0.9169	37.89	94	0.82	32.83
LogitNorm + RefCal2026.05		33.93	0.9067	32.61	3	1.28	29.55
NLL (CE) + RefCal2026.05		32.97	0.9172	16.93	73	0.71	16.79
CE+TS + RefCal2026.05		32.97	0.9254	2.17	59	0.69	2.18
CRL + RefCal2026.05		32.96	0.9172	16.94	73	0.71	16.79
AdaFocal + RefCal2026.05		32.43	0.9261	7.78	68	0.7	7.78
FL+ MDCA + RefCal2026.05		32.25	0.9198	10.98	69	0.7	10.96
MMCE + RefCal2026.05		32.22	0.9321	2	65	0.76	2
AUCM + RefCal2026.05		31.36	0.9295	29.47	8	1.34	27.36
LS + RefCal2026.05		30.01	0.9175	14.86	56	0.93	14.78
MbLS + RefCal2026.05		29.95	0.9211	9.3	62	0.85	9.29
LogitNorm2026.05		25.56	0.8434	24.03	5	1.01	23.18
NLL (CE)2026.05		24.81	0.8408	24.73	80	0.79	23.84
CE+TS2026.05		24.81	0.8509	5.3	57	0.62	5.22
MMCE2026.05		23.54	0.8415	24.85	77	0.78	23.9
LS2026.05		23.53	0.8008	3.36	35	0.49	3.32
MbLS2026.05		23.47	0.008	19.12	64	0.66	18.97
CRL2026.05		23.34	0.8662	7.21	62	0.64	7.11
FL+ MDCA2026.05		22.83	0.8503	11.9	69	0.72	11.89
AUCM2026.05		19.02	0.8371	5.75	68	0.68	5.57