Image Classification on Oxford Pets (Accuracy and Calibration Metrics)
93.9AccuracyBioMedVR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| BioMedVREvaluation Protocol=few-shot, Backbone=ViT-B/162026.06 | 93.9 | — | — | — | |
| AttrVREvaluation Protocol=few-shot, Backbone=ViT-B/162026.06 | 93.1 | — | — | — | |
| AREvaluation Protocol=few-shot, Backbone=ViT-B/162026.06 | 92.6 | — | — | — | |
| VPEvaluation Protocol=few-shot, Backbone=ViT-B/162026.06 | 91.3 | — | — | — | |
| SV-EnsembleBackbone=DINOv2 ViT-S/14, Evaluation Protocol=Fine-tuning, Training Duration=~200 iterations, Number of Ensemble Members=42026.01 | 90.1 | 2.2 | 0.3 | 0.15 | |
| Deep EnsembleBackbone=DINOv2 ViT-S/14, Evaluation Protocol=Fine-tuning, Training Duration=~200 iterations, Number of Ensemble Members=42026.01 | 89.2 | 13.3 | 0.43 | 0.17 | |
| BioMedVREvaluation Protocol=zero-shot, Backbone=ViT-B/162026.06 | 88.4 | — | — | — | |
| MG-CLIP2026.03 | 88.2 | — | — | — | |
| SeGP-CLTSGR=true2026.03 | 87.7 | — | — | — | |
| zero-shot CLIPmode=zero-shot2026.03 | 87.6 | — | — | — | |
| CLIPYear=20212026.04 | 87.35 | — | — | — | |
| Single w/ SVFBackbone=DINOv2 ViT-S/14, Evaluation Protocol=Fine-tuning, Training Duration=~200 iterations2026.01 | 87.2 | 1.9 | 0.4 | 0.19 | |
| OursModel Configuration=5 trees2026.04 | 86.32 | — | — | — | |
| LoRA-EnsembleBackbone=DINOv2 ViT-S/14, Evaluation Protocol=Fine-tuning, Training Duration=~200 iterations, Number of Ensemble Members=42026.01 | 86.1 | 9 | 0.49 | 0.22 | |
| SPU2026.03 | 85.9 | — | — | — | |
| AoSYear=20252026.04 | 85.8 | — | — | — | |
| OursModel Configuration=1 tree2026.04 | 85.63 | — | — | — | |
| SeGP-CLTSGR=false2026.03 | 85.2 | — | — | — | |
| FAREYear=20242026.04 | 84.64 | — | — | — | |
| MoE-Adapter2026.03 | 84.1 | — | — | — | |
| PMG-FTYear=20242026.04 | 82.91 | — | — | — | |
| SingleBackbone=DINOv2 ViT-S/14, Evaluation Protocol=Fine-tuning, Training Duration=~200 iterations2026.01 | 81.4 | 8.6 | 0.61 | 0.27 | |
| TeCoAYear=20232026.04 | 79.42 | — | — | — | |
| CLIPEvaluation Protocol=zero-shot, Backbone=ViT-B/162026.06 | 77.6 | — | — | — | |
| Batch EnsembleBackbone=DINOv2 ViT-S/14, Evaluation Protocol=Fine-tuning, Training Duration=~200 iterations, Number of Ensemble Members=42026.01 | 70.5 | 48.7 | 1.75 | 0.69 | |
| RAPF2026.03 | 28.6 | — | — | — | |
| MC DropoutBackbone=DINOv2 ViT-S/14, Evaluation Protocol=Fine-tuning, Training Duration=~200 iterations2026.01 | 25.9 | 6.1 | 2.53 | 0.85 | |
| PROOF2026.03 | 22.8 | — | — | — |