Video-to-Text Retrieval on MSRVTT 1kA severity degree 4
35.1Gaussian Noise Score (Severity 4)HAT-VTR
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| HAT-VTRBackbone=Xpool2026.02 | 35.1 | 30.1 | 43.7 | 32.1 | 41.1 | 42.5 | 37.4 | 23 | 37.4 | 21.2 | 37.1 | 36.8 | 34.8 | |
| HAT-VTRBackbone=CLIP4Clip2026.02 | 31.8 | 29.1 | 39.6 | 29.2 | 36.9 | 39.2 | 35.2 | 21.9 | 34.4 | 21.1 | 31.6 | 35.9 | 32.2 | |
| EATABackbone=Xpool2026.02 | 29.1 | 2.1 | 39.2 | 23.7 | 37.4 | 41.4 | 33.3 | 16.3 | 33.6 | 16.9 | 34.1 | 35.7 | 28.6 | |
| TCRBackbone=Xpool2026.02 | 27.3 | 21.4 | 39.2 | 24.7 | 38 | 40.6 | 33.1 | 15.1 | 34.6 | 16.3 | 34 | 35.7 | 30 | |
| TCRBackbone=CLIP4Clip2026.02 | 26.6 | 17.3 | 35.5 | 22.7 | 33.6 | 37.6 | 31.4 | 15.4 | 28 | 15.3 | 26.7 | 34.1 | 27 | |
| EATABackbone=CLIP4Clip2026.02 | 25.1 | 1.9 | 34.4 | 20.7 | 32.1 | 38.6 | 31.5 | 13 | 28.6 | 13.3 | 26.8 | 33.4 | 25 | |
| SARBackbone=Xpool2026.02 | 25 | 8.2 | 37.6 | 21.4 | 35.4 | 41 | 31.4 | 14.7 | 33.5 | 12.6 | 34.2 | 35.2 | 27.5 | |
| TentBackbone=Xpool2026.02 | 24.6 | 7 | 37.4 | 21.2 | 35.7 | 41.1 | 31.2 | 14.6 | 33.7 | 12 | 33.8 | 35.1 | 27.3 | |
| XpoolBackbone=Xpool, Adaptation=None2026.02 | 21.6 | 11.1 | 35.6 | 19.8 | 34.4 | 41 | 29 | 12.4 | 32.9 | 12.3 | 33.7 | 35.3 | 26.6 | |
| SARBackbone=CLIP4Clip2026.02 | 21.1 | 5.6 | 32.6 | 17.5 | 29.7 | 37.8 | 30.6 | 10.5 | 26.9 | 10.6 | 26.7 | 33.7 | 23.6 | |
| READBackbone=Xpool2026.02 | 20.8 | 13.2 | 34.6 | 19.2 | 33.5 | 40.6 | 27.7 | 11.4 | 32.5 | 13.3 | 33.5 | 35.2 | 26.3 | |
| TentBackbone=CLIP4Clip2026.02 | 19 | 4.6 | 33.1 | 16.6 | 30 | 37.8 | 30.3 | 9.6 | 26.2 | 9.6 | 26.5 | 34 | 23.1 | |
| READBackbone=CLIP4Clip2026.02 | 18.7 | 10.7 | 31.2 | 16.1 | 27.3 | 37.3 | 28 | 9.7 | 24.8 | 12 | 27.2 | 33.9 | 23.1 | |
| CLIP4ClipBackbone=CLIP4Clip, Adaptation=None2026.02 | 17.8 | 8.4 | 32 | 16.4 | 28.9 | 37.5 | 29.3 | 9.9 | 25.8 | 10.9 | 27 | 34.2 | 23.2 |