Image-to-Text Retrieval on AIC-ICC
61.6R@1Wukong_ViT-L
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Wukong_ViT-LBackbone=ViT-L, Evaluation Protocol=Fine-tuned2022.02 | 61.6 | 80.5 | 86.1 | 71.6 | — | |
| CLIP_ViT-LBackbone=ViT-L, Evaluation Protocol=Fine-tuned2022.02 | 59.1 | 79.5 | 85.2 | 69.9 | — | |
| FILIP_ViT-LBackbone=ViT-L, Evaluation Protocol=Fine-tuned2022.02 | 54.1 | 75.8 | 82.8 | 67.4 | — | |
| FILIP_Swin-LBackbone=Swin-L, Evaluation Protocol=Fine-tuned2022.02 | 53.1 | 74.8 | 82 | 65.2 | — | |
| Wukong_Swin-LBackbone=Swin-L, Evaluation Protocol=Fine-tuned2022.02 | 50.9 | 73.6 | 81.5 | 63.7 | — | |
| CLIP_ViT-BBackbone=ViT-B, Evaluation Protocol=Fine-tuned2022.02 | 50.5 | 73 | 80.2 | 63.1 | — | |
| CLIP_Swin-LBackbone=Swin-L, Evaluation Protocol=Fine-tuned2022.02 | 50.5 | 73.5 | 81.2 | 63 | — | |
| Wukong_ViT-BBackbone=ViT-B, Evaluation Protocol=Fine-tuned2022.02 | 47.5 | 70.6 | 78.6 | 57 | — | |
| WenLan 2.0Evaluation Protocol=Fine-tuned2022.02 | 45.6 | 68 | 76.3 | 58.7 | — | |
| FILIP_ViT-BBackbone=ViT-B, Evaluation Protocol=Fine-tuned2022.02 | 42.5 | 67.2 | 76 | 57.6 | — | |
| ERNIE-ViL 2.0Setting=Zero-shot2022.09 | 33.7 | 52.1 | 60 | — | 40.6 | |
| R2D2Setting=Zero-shot2022.09 | 30.7 | 47.2 | 52.9 | — | 34.5 | |
| R2D2Backbone=ViT-L, Zero-shot=true2022.05 | 30.7 | 47.2 | 52.9 | — | — | |
| FILIPBackbone=ViT-L, Zero-shot=true2022.05 | 20.6 | 37 | 45.4 | — | — | |
| BriVLSetting=Zero-shot2022.09 | 20.3 | 37 | 45.6 | — | 31.1 | |
| WukongBackbone=ViT-L, Zero-shot=true2022.05 | 18.2 | 34.5 | 42.4 | — | — | |
| CLIPBackbone=ViT-L, Zero-shot=true2022.05 | 16.8 | 32 | 39.8 | — | — |