Video Situation Recognition on VidSitu
76.24CIDErSRL-CLIP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SRL-CLIPBackbone=ViT L/142024.01 | 76.24 | 52.36 | 83.91 | |
| SRL-CLIPBackbone=ViT B/162024.01 | 72.11 | 48.78 | 81.95 | |
| SRL-CLIPBackbone=ViT B/322024.01 | 71.5 | 45.88 | 80.66 | |
| VideoWhisperer2024.01 | 68.54 | 45.06 | 75.59 | |
| CLIPBackbone=ViT L/142024.01 | 60.79 | 46.57 | 78.73 | |
| Slow-D+TxE+TxD2024.01 | 60.34 | — | — | |
| CLIPBackbone=ViT B/322024.01 | 55.14 | 44.68 | 79.79 | |
| CLIPBackbone=ViT B/162024.01 | 54.25 | 45.83 | 80.12 | |
| VidSitu2024.01 | 46.01 | 46.79 | 75.9 |