Image Classification on SVHN (test) (Accuracy and Optimization Steps)
95.19Test AccuracyDASH
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DASHBackbone=ResNet-18, Optimizer=SAM2024.10 | 95.19 | 14,467 | 91.67 | 8,613 | |
| DASHBackbone=VGG-16, Optimizer=SAM2024.10 | 94.84 | 8,836 | 92.05 | 5,769 | |
| S&PBackbone=VGG-16, Optimizer=SAM2024.10 | 94.83 | 8,289 | 91.94 | 4,979 | |
| DASHBackbone=VGG-16, Optimizer=SGD2024.10 | 94.72 | 12,668 | 91.84 | 6,844 | |
| S&PBackbone=VGG-16, Optimizer=SGD2024.10 | 94.58 | 9,853 | 91.82 | 5,979 | |
| S&PBackbone=ResNet-18, Optimizer=SAM2024.10 | 94.27 | 11,573 | 89.35 | 5,548 | |
| Random InitBackbone=VGG-16, Optimizer=SAM2024.10 | 93.88 | 12,355 | 90.53 | 7,379 | |
| DASHBackbone=ResNet-18, Optimizer=SGD2024.10 | 93.67 | 5,161 | 89.59 | 2,619 | |
| Random InitBackbone=VGG-16, Optimizer=SGD2024.10 | 93.65 | 16,187 | 90.43 | 8,617 | |
| Warm ReMBackbone=VGG-16, Optimizer=SAM2024.10 | 93.24 | 1,016 | 89.83 | 935 | |
| Warm InitBackbone=VGG-16, Optimizer=SAM2024.10 | 93.08 | 938 | 89.8 | 959 | |
| Warm ReMBackbone=VGG-16, Optimizer=SGD2024.10 | 92.85 | 1,329 | 89.64 | 1,128 | |
| S&PBackbone=ResNet-18, Optimizer=SGD2024.10 | 92.67 | 3,597 | 87.35 | 1,858 | |
| Warm InitBackbone=VGG-16, Optimizer=SGD2024.10 | 92.67 | 1,485 | 89.61 | 1,122 | |
| Random InitBackbone=ResNet-18, Optimizer=SAM2024.10 | 89.84 | 10,869 | 83.31 | 5,546 | |
| Warm InitBackbone=ResNet-18, Optimizer=SAM2024.10 | 88.85 | 1,329 | 81.16 | 993 | |
| Warm ReMBackbone=ResNet-18, Optimizer=SAM2024.10 | 88.75 | 1,485 | 81.03 | 1,006 | |
| Random InitBackbone=ResNet-18, Optimizer=SGD2024.10 | 86.27 | 5,552 | 78.01 | 3,099 | |
| Warm InitBackbone=ResNet-18, Optimizer=SGD2024.10 | 84.01 | 938 | 75.37 | 642 | |
| Warm ReMBackbone=ResNet-18, Optimizer=SGD2024.10 | 83.85 | 782 | 75.41 | 640 | |
| Random InitBackbone=MLP, Optimizer=SAM2024.10 | 83.68 | 36,206 | 78.18 | 18,278 | |
| DASHBackbone=MLP, Optimizer=SAM2024.10 | 83.02 | 15,405 | 77.89 | 8,956 | |
| Random InitBackbone=MLP, Optimizer=SGD2024.10 | 82.92 | 31,768 | 77.19 | 19,861 | |
| S&PBackbone=MLP, Optimizer=SAM2024.10 | 82.81 | 16,734 | 77.94 | 9,802 | |
| DASHBackbone=MLP, Optimizer=SGD2024.10 | 82.3 | 25,571 | 76.77 | 21,092 | |
| Warm InitBackbone=MLP, Optimizer=SAM2024.10 | 82.29 | 2,893 | 77.55 | 2,510 | |
| Warm ReMBackbone=MLP, Optimizer=SAM2024.10 | 82.25 | 3,128 | 77.46 | 2,626 | |
| S&PBackbone=MLP, Optimizer=SGD2024.10 | 82.07 | 28,621 | 77 | 16,530 | |
| Warm ReMBackbone=MLP, Optimizer=SGD2024.10 | 81.21 | 4,398 | 76.53 | 6,147 | |
| Warm InitBackbone=MLP, Optimizer=SGD2024.10 | 81.17 | 4,789 | 76.51 | 7,317 |