Image Classification on CIFAR-10 (test) (Standard Performance & Training Metrics)
95.47Accuracy (Test, Last)Random Init
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Random InitBackbone=ResNet-18, Optimizer=SAM2024.10 | 95.47 | 47,832 | 89.24 | 23,751 | |
| DASHBackbone=ResNet-18, Optimizer=SAM2024.10 | 95.06 | 48,223 | 89.34 | 22,233 | |
| Warm ReMBackbone=ResNet-18, Optimizer=SAM2024.10 | 95 | 47,962 | 88.56 | 21,786 | |
| Warm InitBackbone=ResNet-18, Optimizer=SAM2024.10 | 94.8 | 47,832 | 88.62 | 21,821 | |
| Random InitBackbone=ResNet-18, Optimizer=SGD2024.10 | 94.73 | 50,439 | 88.77 | 24,826 | |
| S&PBackbone=ResNet-18, Optimizer=SAM2024.10 | 94.73 | 48,353 | 89.27 | 22,805 | |
| Warm ReMBackbone=ResNet-18, Optimizer=SGD2024.10 | 94.56 | 51,612 | 88.2 | 23,775 | |
| Warm InitBackbone=ResNet-18, Optimizer=SGD2024.10 | 94.35 | 51,221 | 87.94 | 23,759 | |
| DASHBackbone=ResNet-18, Optimizer=SGD2024.10 | 94.25 | 51,872 | 88.65 | 24,264 | |
| S&PBackbone=ResNet-18, Optimizer=SGD2024.10 | 94.15 | 51,351 | 88.38 | 25,369 |