Image Classification on CIFAR-100 (test) (Symmetric Noise Accuracy 20/50/80/90%) benchmark leaderboard · SOTA2 Research

Image Classification on CIFAR-100 (test) (Symmetric Noise Accuracy 20/50/80/90%)

77.3Accuracy (Symmetric 20%)DivideMix

Evaluation Results

Method	Links
DivideMix2026.03		77.3	74.6	60.2	31.5
LossModel2026.03		73.9	66.1	48.2	24.3
JEC2026.03		73.1	64.9	50.1	16.2
PENCIL2026.03		69.4	57.5	31.1	15.3
Coteaching2026.03		65.6	51.8	27.9	13.7
CE2026.03		62	46.7	19.9	10.1
Forward2026.03		61.5	46.6	19.9	10.2
FEC2026.03		60.5	58.6	52.7	44.1
Mixup2026.03		57.3	30.8	14.6	—