Video Event Prediction on VLEP
72AccuracyVidIL 4-shot
Evaluation Results
| Method | Links | |
|---|---|---|
| VidIL 4-shotNumber of Frames=12, Speech Input=true2023.05 | 72 | |
| SEVILA†Number of Frames=32 -> 4, Keyframe Selection=zero-shot LOCALIZER2023.05 | 69 | |
| SEVILANumber of Frames=32 -> 4, Keyframe Selection=refined LOCALIZER2023.05 | 68.9 | |
| BLIP-2 concat (ANSWERER)Number of Frames=4, Protocol=feature concatenation2023.05 | 68.6 | |
| BLIP-2 votingNumber of Frames=4, Protocol=majority voting2023.05 | 67 | |
| SeViLA†Number of Frames=32 -> 4, Zero-shot=true2023.05 | 64.4 | |
| BLIP-2concat (Answerer)Number of Frames=4, Zero-shot=true2023.05 | 64 | |
| InternVideoNumber of Frames=8, Tested_by_authors=true2023.05 | 63.9 | |
| BLIP-2votingNumber of Frames=4, Zero-shot=true2023.05 | 63.8 | |
| InternVideo*Number of Frames=8, Zero-shot=true2023.05 | 58.7 |