Multimodal Retrieval on MOSI
16.7Recall@10 (Q: M23, T: M1)ConFu
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ConFu2025.11 | 16.7 | 21 | 16.4 | 19.2 | 21 | 21.6 | 16.1 | 23.5 | 20.5 | 19.5 | 19.6 | |
| Symile2025.11 | 16.3 | — | — | — | — | 18.1 | — | — | 17.1 | — | 17.2 | |
| GRAM2025.11 | 16.3 | 6 | 10.3 | 7.7 | 0.3 | 7.9 | 12.2 | 0.2 | 12 | 6.1 | 12.1 | |
| TRIANGLE2025.11 | 8.3 | — | — | — | — | 4.9 | — | — | 8.8 | — | 7.3 | |
| Trimodal CLIP2025.11 | — | 22.9 | 20.5 | 23.5 | 23.3 | — | 20.9 | 24.3 | — | 22.6 | — |