Medical Image Classification on PAD-UFES-20
76.3AccuracyHyperAdapt
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| HyperAdaptBackbone=Swin-T2026.01 | 76.3 | 78.8 | 69.1 | 72.7 | |
| FairQuantizeBackbone=Swin-T2026.01 | 75.9 | 74.6 | 67.8 | 69.7 | |
| VanillaBackbone=Swin-T2026.01 | 75.4 | 75 | 66.2 | 69.2 | |
| GroupModelBackbone=Swin-T2026.01 | 75.4 | 72.9 | 67.3 | 69.4 | |
| FairAdaBNBackbone=Swin-T2026.01 | 74.8 | 72.3 | 67.9 | 68.5 | |
| GroupAdaptBackbone=Swin-T2026.01 | 74.6 | 75.3 | 65.9 | 69.4 | |
| HEALBackbone=Swin-T2026.01 | 73.5 | 75.1 | 68.4 | 70.6 | |
| DAFTBackbone=Swin-T2026.01 | 73.3 | 72.1 | 65.6 | 68.1 | |
| HyperAdaptBackbone=ResNet-502026.01 | 72.6 | 68.1 | 62.5 | 64.4 | |
| FairQuantizeBackbone=ResNet-502026.01 | 70.4 | 66.8 | 57.1 | 59.7 | |
| GroupModelBackbone=ResNet-502026.01 | 68.7 | 64.9 | 58.1 | 59.5 | |
| FairAdaBNBackbone=ResNet-502026.01 | 68 | 62.7 | 56.8 | 58.8 | |
| GroupAdaptBackbone=ResNet-502026.01 | 67.6 | 62.8 | 58.3 | 60 | |
| HEALBackbone=ResNet-502026.01 | 67.2 | 61.5 | 58 | 59.4 | |
| VanillaBackbone=ResNet-502026.01 | 66.1 | 64.1 | 57.4 | 59.6 | |
| DAFTBackbone=ResNet-502026.01 | 65.4 | 64.5 | 58.4 | 59.8 | |
| ResNet-18Backbone=ResNet-182026.03 | 55.49 | — | — | — | |
| Learnable PEPositional Encoding=Learnable2026.03 | 50.58 | — | — | — | |
| No PEPositional Encoding=None2026.03 | 50.29 | — | — | — | |
| RoPEPositional Encoding=RoPE2026.03 | 47.69 | — | — | — |