Binary Classification on NICO++ Car vs. Bus
88.9AccuracyNative Backbone
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Native BackboneBackbone=SigLIP 2 ViT-B/16, Evaluation Protocol=Zero-Shot2026.05 | 88.9 | 61.7 | |
| BAPBackbone=SigLIP 2 ViT-B/16, Evaluation Protocol=Linear Probing2026.05 | 88.2 | 70.3 | |
| BAPBackbone=SigLIP 2 ViT-B/16, Evaluation Protocol=Zero-Shot2026.05 | 87.4 | 71.3 | |
| DIALBackbone=SigLIP 2 ViT-B/16, Evaluation Protocol=Zero-Shot2026.05 | 87.1 | 64.9 | |
| BAP (LP)Backbone=CLIP ConvNeXt-W, Evaluation Protocol=Linear Probing2026.05 | 87 | 81.6 | |
| Data-Matched ControlBackbone=SigLIP 2 ViT-B/16, Evaluation Protocol=Trained Evaluation2026.05 | 86.9 | 61.8 | |
| Native BackboneBackbone=SigLIP 2 ViT-B/16, Evaluation Protocol=Linear Probing2026.05 | 85.2 | 38.6 | |
| Data-Matched ControlBackbone=CLIP ConvNeXt-W, Evaluation Protocol=Linear Probing2026.05 | 84.5 | 63.5 | |
| Native Backbone (ZS)Backbone=CLIP ConvNeXt-W, Evaluation Protocol=Zero-shot2026.05 | 82.8 | 58.2 | |
| BAP (ZS)Backbone=CLIP ConvNeXt-W, Evaluation Protocol=Zero-shot2026.05 | 81.6 | 74.5 | |
| Native Backbone (LP)Backbone=CLIP ConvNeXt-W, Evaluation Protocol=Linear Probing2026.05 | 81.1 | 43.9 | |
| DIALBackbone=CLIP ConvNeXt-W, Evaluation Protocol=Zero-shot2026.05 | 77.3 | 64.2 |