Video-to-text retrieval on DiDeMo (Corruption Robustness)
20.32R@1 (Gaussian)HAT-VTR
Evaluation results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| HAT-VTRBase Model=CLIP4Clip, Query Shift=Ours2026.02 | 20.32 | 19.12 | 35.26 | 25 | 25.2 | 37.55 | 29.48 | 16.73 | 22.91 | 8.17 | 5.58 | 31.97 | 23.11 | |
| HAT-VTRBase Model=Xpool, Query Shift=Ours2026.02 | 20.32 | 22.21 | 37.75 | 27.69 | 30.78 | 39.74 | 33.17 | 17.53 | 26.39 | 11.55 | 33.96 | 33.37 | 27.87 | |
| TCRBase Model=Xpool, Query Shift=TCR2026.02 | 16.93 | 14.84 | 32.37 | 23.21 | 20.02 | 38.05 | 27.79 | 10.76 | 23.61 | 7.57 | 30.28 | 30.68 | 23.01 | |
| EATABase Model=Xpool, Query Shift=EATA2026.02 | 13.75 | 9.96 | 26.59 | 12.75 | 19.42 | 37.75 | 26.89 | 7.47 | 21.02 | 6.47 | 31.47 | 31.08 | 20.39 | |
| TCRBase Model=CLIP4Clip, Query Shift=TCR2026.02 | 13.15 | 14.54 | 27.29 | 18.73 | 18.03 | 33.76 | 23.8 | 10.76 | 16.33 | 6.77 | 2.79 | 29.18 | 17.93 | |
| READBase Model=Xpool, Query Shift=READ2026.02 | 10.16 | 10.76 | 27.19 | 15.84 | 14.34 | 36.85 | 22.91 | 6.57 | 20.32 | 7.97 | 29.58 | 31.08 | 19.46 | |
| SARBase Model=Xpool, Query Shift=SAR2026.02 | 9.86 | 10.16 | 24.5 | 16.14 | 16.73 | 36.95 | 25.5 | 6.08 | 18.23 | 7.97 | 30.38 | 31.08 | 19.47 | |
| XpoolBase Model=Xpool, Query Shift=None2026.02 | 9.36 | 10.56 | 25.2 | 15.64 | 15.24 | 37.25 | 24.2 | 6.37 | 18.73 | 7.87 | 29.88 | 30.98 | 19.27 | |
| TentBase Model=Xpool, Query Shift=Tent2026.02 | 8.67 | 9.16 | 23.51 | 15.64 | 16.73 | 36.95 | 25.8 | 5.98 | 15.24 | 7.57 | 30.38 | 30.98 | 18.88 | |
| READBase Model=CLIP4Clip, Query Shift=READ2026.02 | 6.87 | 7.67 | 22.91 | 12.65 | 10.66 | 32.67 | 20.02 | 5.88 | 14.24 | 4.78 | 1.79 | 29.68 | 14.15 | |
| CLIP4ClipBase Model=CLIP4Clip, Query Shift=None2026.02 | 6.08 | 6.67 | 21.61 | 12.15 | 11.16 | 32.87 | 22.21 | 5.88 | 13.65 | 4.88 | 1.69 | 30.18 | 14.09 | |
| EATABase Model=CLIP4Clip, Query Shift=EATA2026.02 | 5.68 | 2.29 | 24.9 | 10.66 | 16.93 | 33.57 | 25.3 | 6.67 | 14.94 | 5.98 | 1.59 | 29.98 | 14.87 | |
| SARBase Model=CLIP4Clip, Query Shift=SAR2026.02 | 4.58 | 4.78 | 22.31 | 11.85 | 12.55 | 33.17 | 23.01 | 5.08 | 13.25 | 5.08 | 1.79 | 30.18 | 13.97 | |
| TentBase Model=CLIP4Clip, Query Shift=Tent2026.02 | 4.28 | 4.38 | 20.72 | 10.86 | 12.75 | 32.87 | 23.21 | 4.48 | 11.25 | 4.58 | 1.49 | 29.88 | 13.4 |