Image Classification on ImageNet 256x256 (Acc%, Robust Acc%)
78.9Accuracy (%)EGC (Guo et al., 2023)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EGC (Guo et al., 2023)Backbone=U-Net, Params=543M, Steps=10002025.10 | 78.9 | 13.56 | |
| Standard AT (Singh et al., 2023)Backbone=ConvNeXt-L-CvSt, Params=198M, Steps=02025.10 | 78.25 | 33.38 | |
| DAT (T = 110)Backbone=ConvNeXt-L-CvSt, Params=198M, Steps=362025.10 | 75.78 | 56.4 | |
| Standard AT (Salman et al., 2020)Backbone=WRN-50-4, Params=223M, Steps=122025.10 | 71.25 | 45.86 | |
| Standard AT (Salman et al., 2020)Backbone=ResNet-50, Params=26M, Steps=132025.10 | 64.91 | 39.96 | |
| DAT (T = 30)Backbone=WRN-50-4, Params=223M, Steps=172025.10 | 64.45 | 45.84 | |
| DAT (T = 15)Backbone=ResNet-50, Params=26M, Steps=142025.10 | 61.31 | 39.96 | |
| DAT (T = 65)Backbone=WRN-50-4, Params=223M, Steps=192025.10 | 58.78 | 40.74 | |
| DAT (T = 30)Backbone=ResNet-50, Params=26M, Steps=142025.10 | 55.96 | 37.14 |