Adversarial Robustness on ImageNet-LT benchmark leaderboard · SOTA2 Research

Adversarial Robustness on ImageNet-LT

30.02Clean AccuracyFGSM-Our

Evaluation Results

Method	Links
FGSM-Our2021.04		30.02	18.5	17.67
FGSM-Our2021.04		25.88	13.49	11.87
FGSM-AT2021.04		25.64	15.32	14.59
FGSM-AT2021.04		21.94	10.88	9.45
FAST-Our2021.04		15.45	11.51	10.31
FAST-AT2021.04		11.36	8.23	7.16
FAST-Our2021.04		10.76	7.28	6.13
FAST-AT2021.04		7.2	4.52	3.76