Text-to-Video Retrieval on MSRVTT 7k
38.6Recall@1All-in-one + MELTR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| All-in-one + MELTRbase_model=All-in-one2023.03 | 38.6 | 74.4 | 84.7 | — | |
| All-in-one2023.03 | 34.4 | 65.4 | 75.8 | — | |
| Violet + MELTRbase_model=Violet2023.03 | 33.6 | 63.7 | 77.8 | 3 | |
| Violet2023.03 | 31.7 | 60.1 | 74.6 | 3 | |
| UniVL + MELTRbase_model=UniVL2023.03 | 28.5 | 55.5 | 67.6 | 4 | |
| ClipBERT2023.03 | 22.2 | 46.8 | 59.9 | 6 | |
| UniVL-Align2023.03 | 21.2 | 49.6 | 63.1 | 6 | |
| UniVL-Joint2023.03 | 20.6 | 49.1 | 62.9 | 6 | |
| TACO2023.03 | 19.2 | 44.7 | 57.2 | 7 | |
| HERO2023.03 | 16.8 | 43.4 | 57.7 | — | |
| HowTo100M2023.03 | 14.9 | 40.2 | 52.8 | 9 | |
| JSFusion2023.03 | 10.2 | 31.2 | 43.2 | 13 | |
| MIL-NCE2023.03 | 9.9 | 24 | 32.4 | 29.5 |