Multimodal Retrieval on Flickr30k MAG benchmark
91.7R@1GOMA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GOMAMethod Category=Our Method, Feature Backbone=Qwen2-VL2026.05 | 91.7 | 99.6 | 99.9 | |
| DGFMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 71.2 | 90.9 | 94.9 | |
| DMGCMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 49.1 | 66.3 | 72.2 | |
| SmartCLIPMethod Category=Pairwise Image-Text Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 44.2 | 71.5 | 80.5 | |
| LGMRecMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 43.8 | 76.6 | 86.4 | |
| LinearMethod Category=Direct Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 40.3 | 67.6 | 77.6 | |
| CLIP-RefineMethod Category=Pairwise Image-Text Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 32 | 56.4 | 66.6 | |
| MMGCNMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 30.8 | 66 | 79.7 | |
| VSE++Method Category=Pairwise Image-Text Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 20.5 | 43 | 54.2 | |
| MGATMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 15.9 | 42 | 56.9 | |
| D2S-VSEMethod Category=Pairwise Image-Text Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 15.4 | 33.5 | 43.2 | |
| GSMNMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 10.5 | 30 | 41.9 | |
| RawMethod Category=Direct Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 4.4 | 11.1 | 15.4 |