Text-based Image Retrieval on MIT-States
15Rank-1 AccuracyTIRG* + EATV+L
Evaluation results
| Method | Links | |
|---|---|---|
| TIRG* + EATV+Lreproduction=true, vision_backbone=EAT-V, language_backbone=EAT-L2021.05 | 15 | |
| TIRG* + EATVreproduction=true, vision_backbone=EAT-V2021.05 | 14.9 | |
| TIRG* + R101vreproduction=true, vision_backbone=ResNet-1012021.05 | 14.3 | |
| TIRG* + EATLreproduction=true, language_backbone=EAT-L2021.05 | 13.4 | |
| TIRG* + TRreproduction=true, language_backbone=Naive Transformer2021.05 | 13.2 | |
| TIRG*reproduction=true2021.05 | 13.1 | |
| TIRG2021.05 | 12.2 |