Image-to-text Retrieval on Flickr30k (1K)
93.4R@1SoftMask++
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SoftMask++#Img=4M, Zero-shot=true2023.04 | 93.4 | 99.3 | 99.8 | |
| TCL#Img=4M, Zero-shot=true2023.04 | 93 | 99.1 | 99.6 | |
| CODIS#Img=4M, Zero-shot=true2023.04 | 91.7 | 99.3 | 99.8 | |
| ALBEF#Img=4M, Zero-shot=true2023.04 | 90.5 | 98.8 | 99.7 | |
| VISTA-LTime (ms)=~40, Fine-tuning=true2022.03 | 89.5 | 98.4 | 99.6 | |
| ALIGN#Img=1.2B, Zero-shot=true2023.04 | 88.6 | 98.7 | 99.7 | |
| VSEinftyFine-tuning=true2022.03 | 88.4 | 98.3 | 99.5 | |
| CLIP#Img=400M, Zero-shot=true2023.04 | 88 | 98.7 | 99.4 | |
| Pixel-BERT-XTime (ms)=~160, Fine-tuning=true2022.03 | 87 | 98.9 | 99.5 | |
| H Xue et al.Fine-tuning=true2022.03 | 87 | 98.4 | 99.5 | |
| ERNIE-VIL-BTime (ms)=~920, Fine-tuning=true2022.03 | 86.7 | 97.8 | 99 | |
| SOHOFine-tuning=true2022.03 | 86.5 | 98.1 | 99.3 | |
| Unicoder-VLTime (ms)=925, Fine-tuning=true2022.03 | 86.2 | 96.3 | 99 | |
| UNITER-BTime (ms)=~900, Fine-tuning=true2022.03 | 85.9 | 97.1 | 98.8 | |
| VISTA-BTime (ms)=~17, Fine-tuning=true2022.03 | 84.8 | 97.4 | 99 | |
| ViLT-BTime (ms)=~15, Fine-tuning=true2022.03 | 83.5 | 96.7 | 98.6 | |
| UNITER-BaseVisual Embed=Region, Time (ms)=~900, Zero-Shot=true2021.02 | 80.7 | 95.7 | 98 | |
| UNITER#Img=4M, Zero-shot=true2023.04 | 80.7 | 95.7 | 98 | |
| SGRAFFine-tuning=true2022.03 | 77.8 | 94.1 | 97.4 | |
| GSMNFine-tuning=true2022.03 | 76.4 | 94.3 | 97.3 | |
| Pixel-BERT-RTime (ms)=~60, Fine-tuning=true2022.03 | 75.7 | 94.7 | 97.1 | |
| IMRAMFine-tuning=true2022.03 | 74.1 | 93 | 96.6 | |
| ViLT-B/32Visual Embed=Linear, Time (ms)=~15, Zero-Shot=true, Pre-training steps=200K2021.02 | 73.2 | 93.6 | 96.5 | |
| ViLT#Img=4M, Zero-shot=true2023.04 | 73.2 | 93.6 | 96.5 | |
| VSRNFine-tuning=true2022.03 | 71.3 | 90.6 | 96 | |
| ImageBERTVisual Embed=Region, Time (ms)=~925, Zero-Shot=true, Additional Pre-training Dataset=10M proprietary2021.02 | 70.7 | 90.2 | 94 | |
| ViLT-B/32Visual Embed=Linear, Time (ms)=~15, Zero-Shot=true2021.02 | 69.7 | 91 | 96 | |
| SCANFine-tuning=true2022.03 | 67.4 | 90.3 | 95.8 | |
| Unicoder-VLVisual Embed=Region, Time (ms)=~925, Zero-Shot=true2021.02 | 64.3 | 85.8 | 92.3 | |
| Vil-BERTTime (ms)=~920, Fine-tuning=true2022.03 | 58.2 | 84.9 | 91.5 |