Text-to-Image Retrieval on FashionGen (test)
70.12R@1FashionSAP
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FashionSAP2023.04 | 70.12 | 91.76 | 96.38 | — | 71.63 | |
| FashionSAPBackbone=ViT-B162023.04 | 69.07 | 89.81 | 94.75 | — | 70.11 | |
| FashionSAPBackbone=ResNet-502023.04 | 64.11 | 88.24 | 94.31 | — | 65.67 | |
| FAME-ViL(ST)Backbone=ViT-B/16, Learning Strategy=STL, Test Protocol=random 1002023.03 | 63.56 | 86.84 | 93.06 | 82.61 | — | |
| FAME-ViLBackbone=ViT-B/16, Learning Strategy=MTL, Test Protocol=random 1002023.03 | 62.86 | 87.38 | 93.52 | 83.14 | — | |
| FashionViLBackbone=ResNet-50, Learning Strategy=STL, Test Protocol=random 1002023.03 | 61.88 | 87.32 | 93.22 | 82.6 | — | |
| FashionViL2023.04 | 61.88 | 87.32 | 93.22 | — | 63.71 | |
| FashionViL(vit)Backbone=ViT-B/16, Learning Strategy=STL, Test Protocol=random 1002023.03 | 60.76 | 86.18 | 92.96 | 81.61 | — | |
| ALBEF2023.04 | 60.52 | 84.99 | 91.45 | — | 62.25 | |
| EI-CLIPLearning Strategy=STL, Test Protocol=random 1002023.03 | 40.06 | 71.99 | 82.9 | 65.02 | — | |
| EI-CLIP2023.04 | 40.06 | 71.99 | 82.9 | — | 39.38 | |
| CommerceMM2023.04 | 39.6 | 61.5 | 72.7 | — | 40.6 | |
| MVLTLearning Strategy=STL, Test Protocol=random 1002023.03 | 34.6 | 78 | 89.5 | 67.25 | — | |
| KaleidoBERTLearning Strategy=STL, Test Protocol=random 1002023.03 | 33.88 | 60.6 | 68.59 | 53.25 | — | |
| KaleidoBERT2023.04 | 33.88 | 60.6 | 68.59 | — | 30.94 | |
| FashionBERTLearning Strategy=STL, Test Protocol=random 1002023.03 | 26.75 | 46.48 | 55.74 | 41.89 | — | |
| FashionBERT2023.04 | 26.75 | 46.48 | 55.74 | — | 25.36 | |
| OSCARLearning Strategy=STL, Test Protocol=random 1002023.03 | 25.1 | 49.14 | 56.68 | 41.92 | — | |
| OSCAR2023.04 | 25.1 | 49.14 | 56.68 | — | 24.25 | |
| Image-BERT2023.04 | 24.78 | 45.2 | 55.9 | — | 23.77 | |
| VL-BERT2023.04 | 22.63 | 36.48 | 48.52 | — | 20.95 | |
| VILBERT2023.04 | 21.12 | 37.23 | 50.11 | — | 21.05 |