Multimodal Retrieval on Movies MAG benchmark
80.3Recall@1GOMA
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GOMAMethod Category=Our Method, Feature Backbone=Qwen2-VL2026.05 | 80.3 | 97.9 | 99 | |
| DGFMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 79.5 | 95.3 | 97.6 | |
| DMGCMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 69.4 | 81.9 | 85.8 | |
| SmartCLIPMethod Category=Pairwise Image-Text Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 62.7 | 83.4 | 87.8 | |
| D2S-VSEMethod Category=Pairwise Image-Text Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 54.5 | 78.3 | 84 | |
| VSE++Method Category=Pairwise Image-Text Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 51.2 | 76.2 | 82.5 | |
| CLIP-RefineMethod Category=Pairwise Image-Text Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 50.5 | 72.9 | 79 | |
| LinearMethod Category=Direct Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 37.7 | 62.3 | 71.3 | |
| LGMRecMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 32.5 | 63.8 | 75.1 | |
| MMGCNMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 14.1 | 40.5 | 55.7 | |
| MGATMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 12.9 | 37.2 | 51.5 | |
| GSMNMethod Category=Multimodal Graph Retrieval & Recent MAG Models, Feature Backbone=Qwen2-VL2026.05 | 10.3 | 29.8 | 42.3 | |
| RawMethod Category=Direct Retrieval Baselines, Feature Backbone=Qwen2-VL2026.05 | 9.8 | 18.6 | 23.3 |