Video Narrating on E-HVC-Bench (test)
14.48SODA cHiVid-Narrator
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| HiVid-NarratorNumber of scene tokens (S)=64, Number of event tokens (E)=322026.01 | 14.48 | 1.45 | 32.01 | 74.25 | |
| Keye-VL(8B)Number of parameters=8B2026.01 | 12.83 | 1.34 | 28.29 | 69.23 | |
| Qwen2.5-VL(7B)Number of parameters=7B2026.01 | 12.27 | 1.12 | 27.74 | 68.83 | |
| InternVL3(8B)Number of parameters=8B2026.01 | 11.98 | 1.04 | 27.38 | 67.75 |