Text-to-Video Retrieval on MSRVTT (MSR) zero-shot
43.3R@1Uncompressed
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| UncompressedPruning Mode=/, Pruning Ratio=/, Backbone=BLIP, Zero-shot=true2026.04 | 43.3 | 65.5 | — | 733.4 | |
| UMT-L#Pairs=17M, Zero-shot=true2023.03 | 42.6 | — | — | — | |
| BT-AdapterPretraining Scale=2M, Backbone=CLIP-L/14, Zero-shot=true2023.09 | 40.9 | 64.7 | 73.5 | — | |
| InternVideo#Pairs=646M, Zero-shot=true2023.03 | 40.7 | — | — | — | |
| UMT-L#Pairs=25M, Zero-shot=true2023.03 | 40.7 | — | — | — | |
| InternVideoPretraining Scale=12.8M, Backbone=CLIP-L/14, Zero-shot=true2023.09 | 40.7 | — | — | — | |
| CoMPPruning Mode=C, Pruning Ratio=0.7, Backbone=BLIP, Zero-shot=true2026.04 | 39.5 | 63.1 | — | 261.1 | |
| MADTPPruning Mode=T, Pruning Ratio=0.7, Backbone=BLIP, Zero-shot=true2026.04 | 38.9 | 62.4 | — | 265.3 | |
| TVTSv2Pretraining Scale=8.5M, Backbone=CLIP-L/14, Zero-shot=true2023.09 | 38.2 | 62.4 | 73.2 | — | |
| ImageBindBackbone=CLIP-L/14, Zero-shot=true2023.09 | 36.8 | 61.8 | 70 | — | |
| UMT-B#Pairs=17M, Zero-shot=true2023.03 | 35.5 | — | — | — | |
| CLIPBackbone=CLIP-L/14, Zero-shot=true2023.09 | 35.4 | 58.8 | 68.1 | — | |
| UMT-B#Pairs=25M, Zero-shot=true2023.03 | 35.2 | — | — | — | |
| OmniVL#Pairs=17M, Zero-shot=true2023.03 | 34.6 | — | — | — | |
| OmniVLPretraining Scale=14M, Backbone=Non-CLIP, Zero-shot=true2023.09 | 34.6 | 58.4 | 66.6 | — | |
| VideoCoCa#Pairs=4.8B, Zero-shot=true2023.03 | 34.3 | — | — | — | |
| VideoCoCaPretraining Scale=100M, Backbone=Non-CLIP, Zero-shot=true2023.09 | 34.3 | 57.8 | 67 | — | |
| Singularity#Pairs=17M, Zero-shot=true2023.03 | 34 | — | — | — | |
| SingularityPretraining Scale=17M, Backbone=Non-CLIP, Zero-shot=true2023.09 | 34 | 56.7 | 66.7 | — | |
| UMT-L#Pairs=5M, Zero-shot=true2023.03 | 33.3 | — | — | — | |
| UMT-LPretraining Scale=5M, Backbone=CLIP-L/14, Zero-shot=true2023.09 | 33.3 | 58.1 | 66.7 | — | |
| VINDLU#Pairs=25M, Zero-shot=true2023.03 | 32 | — | — | — | |
| CLIP4Clip#Pairs=400M, Zero-shot=true2023.03 | 30.6 | — | — | — | |
| HiTeAPretraining Scale=5M, Backbone=Non-CLIP, Zero-shot=true2023.09 | 29.9 | 54.2 | 62.9 | — | |
| UMT-B#Pairs=5M, Zero-shot=true2023.03 | 29.6 | — | — | — | |
| CloverPretraining Scale=5M, Backbone=Non-CLIP, Zero-shot=true2023.09 | 26.4 | 49.5 | 60 | — | |
| VIOLET#Pairs=138M, Zero-shot=true2023.03 | 25.9 | — | — | — | |
| FrozenPretraining Scale=5M, Backbone=Non-CLIP, Zero-shot=true2023.09 | 24.7 | 46.9 | 57.2 | — | |
| UPopPruning Mode=P, Pruning Ratio=0.7, Backbone=BLIP, Zero-shot=true2026.04 | 24 | 42.2 | — | 256.2 | |
| Frozen#Pairs=5M, Zero-shot=true2023.03 | 18.7 | — | — | — |