Text-to-Video Retrieval on ActivityNet LSMDC → ActivityNet
31.4R@1HAT-VTR
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| HAT-VTRBackbone=CLIP4Clip2026.02 | 31.4 | 58.71 | 70.55 | |
| HAT-VTRBackbone=Xpool2026.02 | 26.8 | 53.39 | 65.87 | |
| EATABackbone=Xpool2026.02 | 23.55 | 49.4 | 63.11 | |
| EATABackbone=CLIP4Clip2026.02 | 22.45 | 46.82 | 60.85 | |
| TCRBackbone=CLIP4Clip2026.02 | 22.05 | 47.14 | 61.93 | |
| SARBackbone=Xpool2026.02 | 22.01 | 47.28 | 60.32 | |
| SARBackbone=CLIP4Clip2026.02 | 21.94 | 47.41 | 60.85 | |
| TCRBackbone=Xpool2026.02 | 21.82 | 47.41 | 60.42 | |
| TentBackbone=CLIP4Clip2026.02 | 21.58 | 47.24 | 60.48 | |
| TentBackbone=Xpool2026.02 | 21.21 | 46.45 | 59.43 | |
| CLIP4ClipBackbone=CLIP4Clip2026.02 | 20.28 | 44.91 | 58.71 | |
| XpoolBackbone=Xpool2026.02 | 19.83 | 44.3 | 56.74 | |
| READBackbone=CLIP4Clip2026.02 | 17.84 | 39.35 | 53.12 | |
| READBackbone=Xpool2026.02 | 17.84 | 39.6 | 51.6 |