Image Classification on CIFAR-10 (test) (Top-1 Accuracy vs Epoch)
92.22Top-1 Accuracy (Epoch 50)LPSGDM
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LPSGDMModel=ResNet-502026.06 | 92.22 | 94.07 | 94.84 | |
| LPSGDMModel=ResNet-182026.06 | 92.12 | 93.68 | 94.81 | |
| AdamWModel=ResNet-182026.06 | 91.78 | 93.09 | 94.09 | |
| AdamWModel=ResNet-502026.06 | 91.65 | 93.45 | 94.57 | |
| LionModel=ResNet-502026.06 | 91.15 | 92.87 | 94.34 | |
| LionModel=ResNet-182026.06 | 90.53 | 91.58 | 94.04 | |
| SGD w/ MomentumModel=ResNet-182026.06 | 88.77 | 91.7 | 93.15 | |
| SGD w/ MomentumModel=ResNet-502026.06 | 87.57 | 90.73 | 92.72 | |
| LPSGDMModel=VGG-112026.06 | 86.15 | 88.83 | 90.73 | |
| SGD w/ MomentumModel=VGG-112026.06 | 85.88 | 88.43 | 90.05 | |
| LionModel=VGG-112026.06 | 85.05 | 83.79 | 89.4 | |
| AdamWModel=VGG-112026.06 | 83.54 | 87.83 | 89.93 |