Main Object Video Captioning on MiraData-9k 1.0 (test)
84.9HalFscore (Hal, Object)HALVA
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| HALVAbase_model=LLaVA-Video2025.12 | 84.9 | 77.2 | 46.7 | 24 | 22.8 | 23.4 | 55 | 60.1 | 49.9 | 23.5 | 47.4 | 29.5 | |
| LLaVA-Videobase_model=LLaVA-Video2025.12 | 83.1 | 75.4 | 43.1 | 23.8 | 24.3 | 24.1 | 53.2 | 58.1 | 46.1 | 24.3 | 46.4 | 30.8 | |
| DeCobase_model=LLaVA-Video2025.12 | 80.9 | 73.5 | 37.7 | 20.7 | 25.3 | 23.2 | 53.6 | 57.2 | 39.5 | 20.9 | 42.7 | 28.1 | |
| SFTbase_model=LLaVA-Video2025.12 | 79.5 | 68.1 | 49.7 | 23.3 | 29 | 26.9 | 44.2 | 55.4 | 54.8 | 21.4 | 55.2 | 28.9 | |
| HACLbase_model=LLaVA-Video2025.12 | 79.4 | 67.9 | 52.3 | 26.3 | 29.6 | 28.9 | 45.6 | 55.8 | 60.8 | 25.7 | 57.5 | 32.5 | |
| SANTAbase_model=LLaVA-Video2025.12 | 77.7 | 65.7 | 53.7 | 27.1 | 31.6 | 30.3 | 43.3 | 54.6 | 61.7 | 26.6 | 59.1 | 33.5 |