Caption Retrieval on Flickr30k
95.5R@1CREM 7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CREM 7BZero-shot=true, Parameters=7B2026.02 | 95.5 | — | — | |
| UniMEZero-shot=true2026.02 | 94.4 | — | — | |
| CREM 2BZero-shot=true, Parameters=2B2026.02 | 91.9 | — | — | |
| VLM2VecZero-shot=true2026.02 | 90.6 | — | — | |
| Discrete-continuous PGNetworks=Faster R-CNN2021.04 | 82.8 | 97.9 | 95.9 | |
| VSRNNetworks=Faster R-CNN2021.04 | 71.3 | 96 | 90.6 | |
| SCANNetworks=Faster R-CNN2021.04 | 67.4 | 95.8 | 90.3 | |
| SCONetworks=ResNet2021.04 | 55.5 | 89.3 | 82 | |
| VSENetworks=ResNet2021.04 | 52.9 | 87.2 | 79.1 | |
| 2wayNNetworks=VGG2021.04 | 49.8 | — | 67.5 | |
| DMENumber of parameters=15M2018.04 | 49.7 | 83.6 | — | |
| CDMENumber of parameters=15M2018.04 | 49 | 83.8 | — | |
| Unweighted DMENumber of parameters=15M2018.04 | 48.9 | 83.7 | — | |
| RRFNetworks=ResNet2021.04 | 47.6 | 87.1 | 77.4 | |
| FastTextNumber of parameters=15M2018.04 | 47.1 | 82.7 | — | |
| NaiveNumber of parameters=32M2018.04 | 46.4 | 82.2 | — | |
| VSE++Backbone=ResNet, Cropping=random, Fine-tuning=false2018.04 | 43.7 | 82.1 | — | |
| SMlstmNetworks=VGG2021.04 | 42.5 | 81.5 | 71.9 | |
| HMlstmNetworks=AlexNet2021.04 | 38.1 | 76.5 | — | |
| ImageNetNumber of parameters=29M2018.04 | 36.6 | 72.2 | — | |
| FVNetworks=VGG2021.04 | 35 | 73.8 | 62 | |
| VQANetworks=VGG2021.04 | 33.9 | 73.8 | 62.5 | |
| DSVANetworks=AlexNet2021.04 | 32.2 | 61.4 | 48.2 |