Text-to-video retrieval on MSVD 10s (test)
39.3R@1MAC
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MACArchitecture=V,T, # Pairs=5.5M, # Frames=82022.12 | 39.3 | 67.5 | 79.4 | |
| OA-TransArchitecture=V,T,Vdet, # Pairs=5.5M, # Frames=82022.12 | 39.1 | 68.4 | 80.3 | |
| MACArchitecture=V,T, # Pairs=5.5M, # Frames=42022.12 | 36.9 | 65.1 | 75.7 | |
| FrozenArchitecture=V,T, # Pairs=5.5M, # Frames=42022.12 | 33.7 | 64.7 | 76.3 | |
| SupportSetArchitecture=V,T,C, # Pairs=120M, # Frames=-2022.12 | 28.4 | 60 | 72.9 | |
| CEArchitecture=V,T,Vexpert, # Pairs=-, # Frames=-2022.12 | 19.8 | 49 | 63.8 |