RBC classification on DSE (five-fold cross-validation)
86.6Weighted F1 ScoreRedDino large
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| RedDino largeProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 86.6 | 60.1 | 86.6 | |
| DinoBloom-LProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 86.2 | 60.7 | 86.2 | |
| RedDino baseProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 85.9 | 57.9 | 86 | |
| DinoBloom-BProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 85.3 | 59.9 | 85.4 | |
| RedDino smallProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 84.9 | 56.5 | 84.9 | |
| DINOv2 baseProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 84.8 | 56.2 | 84.8 | |
| DINOv2 largeProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 84.7 | 54.4 | 84.9 | |
| DinoBloom-SProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 84.4 | 57.5 | 84.3 | |
| DINOv2 smallProtocol=Linear probing, Validation=five-fold cross-validation2025.08 | 83.6 | 54.4 | 83.6 | |
| ResNet50Protocol=Linear probing, Validation=five-fold cross-validation2025.08 | 83.2 | 51.5 | 83.2 |