Image Classification on LAB → LAB
96.01Average AccuracyInfoNCE (I+P→I+P)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| InfoNCE (I+P→I+P)Model=ViT Tiny2026.03 | 96.01 | 0.56 | |
| InfoNCE (I+P→I+P)Model=EfficientNet B02026.03 | 95.06 | 0.58 | |
| InfoNCE (I+P→I+P)Model=ConvNeXt Femto2026.03 | 94.81 | 0.9 | |
| InfoNCE (I→I)Model=ViT Tiny2026.03 | 92.79 | 0.55 | |
| InfoNCE (I→I)Model=EfficientNet B02026.03 | 91.62 | 1.05 | |
| InfoNCE (I→I)Model=ConvNeXt Femto2026.03 | 90.01 | 0.98 | |
| DINOModel=EfficientNet B02026.03 | 79.67 | 1.42 | |
| DINOModel=ConvNeXt Femto2026.03 | 62.85 | 2.82 | |
| DINOModel=ViT Tiny2026.03 | 59.54 | 2.04 |