Image Classification on CIFAR-10 (Accuracy, FLOPs, and Inference Speed)
92.1AccuracyMOE (Ours)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MOE (Ours)2026.06 | 92.1 | — | — | |
| InceptionV3 (Ours)Total Parameters=3.23M2026.06 | 90.3 | 10.86 | 0.037 | |
| VGG16 (Original)Total Parameters=21.15M2026.06 | 89.8 | — | 0.162 | |
| Resnet18 (Original)Total Parameters=11.2M2026.06 | 88.8 | — | 0.042 | |
| VGG16 (Ours)Total Parameters=3.3M2026.06 | 88.6 | 10.71 | 0.041 | |
| Densenet121 (Ours)Total Parameters=0.648M2026.06 | 88.6 | 10.51 | 0.037 | |
| Densenet121 (Original)Total Parameters=6.87M2026.06 | 87.6 | — | 0.052 | |
| InceptionV3 (Original)Total Parameters=35.37M2026.06 | 86.6 | — | 0.068 | |
| Resnet18 (Ours)Total Parameters=1.01M2026.06 | 85.8 | 10.6 | 0.018 |