Image Classification on CIFAR-10 (test) (Accuracy and Training Efficiency)
88.21AccuracyBaseline
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| BaselineBackbone=EfficientNet-B0, Training Protocol=From Scratch2026.04 | 88.21 | 30 | — | — | |
| DBPDBackbone=EfficientNet-B0, Training Protocol=From Scratch2026.04 | 87.82 | 5.86 | — | — | |
| Adaptive-αBackbone=EfficientNet-B0, Training Protocol=From Scratch2026.04 | 85.69 | 7.45 | -2.52 | 4.72 | |
| DBPDBackbone=ResNet-50, Training Protocol=From Scratch2026.04 | 85.61 | 7.08 | — | — | |
| BaselineBackbone=MobileNet-V2, Training Protocol=From Scratch2026.04 | 85.49 | 30 | — | — | |
| DBPDBackbone=MobileNet-V2, Training Protocol=From Scratch2026.04 | 85.39 | 6.42 | — | — | |
| Adaptive-TBackbone=EfficientNet-B0, Training Protocol=From Scratch2026.04 | 85.28 | 6.36 | -2.52 | 4.72 | |
| BaselineBackbone=ResNet-50, Training Protocol=From Scratch2026.04 | 84.85 | 30 | — | — | |
| Adaptive-αBackbone=ResNet-50, Training Protocol=From Scratch2026.04 | 83.82 | 8.48 | -2.21 | 4.81 | |
| Adaptive-αBackbone=MobileNet-V2, Training Protocol=From Scratch2026.04 | 83.08 | 8.76 | -2.41 | 4.89 | |
| Adaptive-TBackbone=ResNet-50, Training Protocol=From Scratch2026.04 | 82.64 | 6.24 | -2.21 | 4.81 | |
| DBPDBackbone=EfficientFormer-L1, Training Protocol=From Scratch2026.04 | 81.17 | 5.82 | — | — | |
| Adaptive-TBackbone=MobileNet-V2, Training Protocol=From Scratch2026.04 | 80.79 | 6.14 | -2.41 | 4.89 | |
| Adaptive-αBackbone=EfficientFormer-L1, Training Protocol=From Scratch2026.04 | 80.75 | 7.21 | 0.43 | 5.15 | |
| Adaptive-TBackbone=EfficientFormer-L1, Training Protocol=From Scratch2026.04 | 80.67 | 6.9 | 0.43 | 5.15 | |
| BaselineBackbone=EfficientFormer-L1, Training Protocol=From Scratch2026.04 | 80.32 | 30 | — | — |