Text-to-Video Retrieval on ActivityNet 180s (test)
34.4R@1MAC
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MACArchitecture=V,T, # Pairs=5.5M, # Frames=82022.12 | 34.4 | 64.8 | 78.3 | |
| MACArchitecture=V,T, # Pairs=5.5M, # Frames=42022.12 | 30.6 | 60.5 | 74.7 | |
| HiTArchitecture=V,T,Vema, # Pairs=120M, # Frames=322022.12 | 29.6 | 60.7 | — | |
| SupportSetArchitecture=V,T,C, # Pairs=120M, # Frames=-2022.12 | 29.2 | 61.6 | — | |
| MMTArchitecture=V,T,C, # Pairs=120M, # Frames=322022.12 | 28.7 | 61.4 | — | |
| HD-VILAArchitecture=V,T,C, # Pairs=103M, # Frames=142022.12 | 28.5 | 57.4 | 94 | |
| FrozenArchitecture=V,T, # Pairs=5.5M, # Frames=42022.12 | 27.6 | 57.4 | 71.9 | |
| TACoArchitecture=V,T,C, # Pairs=120M, # Frames=482022.12 | 25.8 | 56.3 | 93.8 | |
| All-in-oneArchitecture=U, # Pairs=122.5M, # Frames=92022.12 | 22.4 | 53.7 | 67.7 | |
| ClipBERTArchitecture=V,T,C, # Pairs=5.6M, # Frames=82022.12 | 21.3 | 49 | 63.5 | |
| CEArchitecture=V,T,Vexpert, # Pairs=-, # Frames=-2022.12 | 18.2 | 47.4 | — |