Image Classification on CIFAR-10 standard (test) (Adversarial Robustness and Efficiency)
88.07AccuracyEED
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| EEDSparsity rate (sr)=50%, Backbone=ResNet18, Params=5.6M2025.04 | 88.07 | 57.83 | 52.35 | 57.92 | 53.12 | 1.64 | |
| ATSparsity rate (sr)=0%, Backbone=ResNet18, Params=11.2M2025.04 | 87.05 | 56.14 | 48.02 | 57.6 | 53.1 | 1 | |
| EEDSparsity rate (sr)=70%, Backbone=ResNet18, Params=3.4M2025.04 | 86.45 | 56.32 | 49.16 | 57.51 | 52.74 | 1.77 | |
| HARPSparsity rate (sr)=50%, Backbone=ResNet18, Params=5.6M2025.04 | 86.32 | 55.29 | 46.93 | 57.9 | 50.41 | 1.34 | |
| EEDSparsity rate (sr)=80%, Backbone=ResNet18, Params=2.2M2025.04 | 86.13 | 55.71 | 48.13 | 57.03 | 51.97 | 1.86 | |
| HYDRASparsity rate (sr)=50%, Backbone=ResNet18, Params=5.6M2025.04 | 85.67 | 54.58 | 47.12 | 56.98 | 50.3 | 1.28 | |
| EEDSparsity rate (sr)=90%, Backbone=ResNet18, Params=1.1M2025.04 | 83.26 | 52.14 | 46.85 | 56.77 | 50.21 | 2.51 | |
| Flying BirdSparsity rate (sr)=50%, Backbone=ResNet18, Params=5.6M2025.04 | 83.02 | 53.18 | 45.13 | 57.33 | 48.02 | 1.24 | |
| R-ADMMSparsity rate (sr)=50%, Backbone=ResNet18, Params=5.6M2025.04 | 82.14 | 54.1 | 46.23 | 56.27 | 45.8 | 1.36 | |
| TwinRepSparsity rate (sr)=50%, Backbone=ResNet18, Params=5.6M2025.04 | 81.45 | 54.13 | 48.45 | 57.82 | 48.27 | 1.3 | |
| MADSparsity rate (sr)=50%, Backbone=ResNet18, Params=5.6M2025.04 | 77.33 | 53.21 | 44.45 | 57.21 | 43.12 | 1.37 | |
| EEDSparsity rate (sr)=95%, Backbone=ResNet18, Params=0.6M2025.04 | 74.36 | 48.09 | 45.69 | 42.54 | 46.22 | 3.27 | |
| RSTSparsity rate (sr)=50%, Backbone=ResNet18, Params=5.6M2025.04 | 72.55 | 49.19 | 42.03 | 42.7 | 38.09 | 1.52 |