Image-Text Matching on EQ-GEBC (EQBEN)
25.19Text ScoreFIBER
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FIBERZero-shot=true2023.03 | 25.19 | 22.66 | 11.08 | |
| METER + EQSIMEQSIM=true, Fine-tuning=Flickr30K2023.03 | 24.2 | 26.02 | 11.69 | |
| FIBER + EQSIMEQSIM=true, Fine-tuning=Flickr30K2023.03 | 21.55 | 26.07 | 10.58 | |
| METERZero-shot=true2023.03 | 20.95 | 18.19 | 6.95 | |
| FIBER + FT (F30K)Fine-tuning=Flickr30K2023.03 | 18.24 | 21.33 | 8.54 | |
| METER + FT (F30K)Fine-tuning=Flickr30K2023.03 | 18.08 | 19.85 | 7.33 | |
| BLIPZero-shot=true2023.03 | 15.87 | 19.79 | 7.27 | |
| ViLTZero-shot=true2023.03 | 14.72 | 16.7 | 5.62 | |
| LXMERTZero-shot=true2023.03 | 14.67 | 12.64 | 4.19 | |
| VILBERTZero-shot=true2023.03 | 14.67 | 12.64 | 4.82 | |
| ALBEFZero-shot=true2023.03 | 13.56 | 19.63 | 5.89 | |
| CLIP (ViT-B/32)Zero-shot=true, Backbone=ViT-B/322023.03 | 12.57 | 20.12 | 4.47 | |
| FLAVAZero-shot=true2023.03 | 12.24 | 16.81 | 5.07 | |
| CLIP (RN-50)Zero-shot=true, Backbone=ResNet-502023.03 | 10.8 | 18.03 | 3.97 |