Image-to-Text Retrieval on MSCOCO EN
58.6R@1AltCLIP
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AltCLIPBackbone=ViT-L, Zero-shot=true, Stage=Contrastive Learning2022.11 | 58.6 | 80.6 | 87.8 | 69.2 | |
| AltCLIPTBackbone=ViT-L, Zero-shot=true, Stage=Teacher Learning2022.11 | 57.2 | 80.5 | 87.5 | 66.1 | |
| CLIPBackbone=ViT-L, Zero-shot=true2022.11 | 56.4 | 79.5 | 86.5 | 65.2 | |
| CN-CLIPBackbone=ViT-L, Zero-shot=true2022.11 | 40.9 | 65.8 | 76.3 | 53.4 | |
| TaiyiBackbone=ViT-L, Zero-shot=true2022.11 | 19.8 | 42.1 | 54.3 | 32.2 |