Image Classification on CIFAR-100 (test) (Accuracy and Mean Corruption Accuracy)
0.8312AccuracySYN+NST+TA (Ours+RE+N+FN)
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SYN+NST+TA (Ours+RE+N+FN)Backbone=WRN-28-4, Training Epochs=6002025.12 | 0.8312 | 75.27 | 0.749 | 64.31 | |
| NoisyMixBackbone=WRN-28-4, Training Epochs=600, Evaluation Protocol=Original implementation2025.12 | 0.8177 | 73.24 | 0.7278 | — | |
| NoisyMixBackbone=WRN-28-4, Training Epochs=600, Evaluation Protocol=Robustbench2025.12 | 0.8116 | — | 0.7206 | — |