List your company

Video-to-Text Retrieval on MSRVTT 1kA (Noise Robustness)

41.6Recall@1 (20% Noise)HAT-VTR
35.8837.36538.8540.335Feb 15, 2026

Evaluation results

MethodLinks
HAT-VTRBackbone=Xpool, Adaptation Method=Ours2026.0241.637.632.32935.1
EATABackbone=Xpool, Adaptation Method=EATA2026.0239.934.929.325.932.5
TentBackbone=Xpool, Adaptation Method=Tent2026.023934.129.223.731.5
SARBackbone=Xpool, Adaptation Method=SAR2026.0238.934.229.323.831.6
TCRBackbone=Xpool, Adaptation Method=TCR2026.0238.634.129.324.331.6
READBackbone=Xpool, Adaptation Method=READ2026.0238.533.728.523.431
XpoolBackbone=Xpool, Adaptation Method=None2026.0238.433.628.823.831.2
HAT-VTRBackbone=CLIP4Clip, Adaptation Method=Ours2026.0237.434.429.327.632.2
CLIP4ClipBackbone=CLIP4Clip, Adaptation Method=None2026.0236.731.127.22129
EATABackbone=CLIP4Clip, Adaptation Method=EATA2026.0236.631.627.723.729.9
SARBackbone=CLIP4Clip, Adaptation Method=SAR2026.0236.431.127.321.429.1
TentBackbone=CLIP4Clip, Adaptation Method=Tent2026.0236.331.127.221.529
READBackbone=CLIP4Clip, Adaptation Method=READ2026.0236.130.726.92128.7
TCRBackbone=CLIP4Clip, Adaptation Method=TCR2026.0236.130.526.521.728.7