Image Classification on ImageNet (Accuracy and Accuracy Gain)
72.7AccuracyInMaP + ZLaP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| InMaP + ZLaPBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.04 | 72.7 | 0.2 | |
| InMaPBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.04 | 72.5 | — | |
| CLIP + ZLaPBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.04 | 69.7 | 0.9 | |
| TPTBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.04 | 69 | — | |
| CLIPBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.04 | 68.8 | — | |
| CLIP-DNBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.04 | 68.3 | — | |
| InMaPBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.04 | 63.8 | — | |
| InMaP + ZLaPBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.04 | 63.8 | 0 | |
| CLIP + ZLaPBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.04 | 61.8 | 1.5 | |
| TPTBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.04 | 60.7 | — | |
| CLIPBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.04 | 60.3 | — | |
| CLIP-DNBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.04 | 60.2 | — |