Video Captioning on E-HVC-Bench 1.0 (test)
14.48SODA_cHiVid-Narrator
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| HiVid-Narratorevaluation_mode=Fine-tuned, training_set=E-HVC-146K, SPA-Compressor=w/, scene_tokens (S)=64, event_tokens (E)=322026.01 | 14.48 | 1.45 | 32.01 | 74.25 | |
| HiVid-Narratorevaluation_mode=Fine-tuned, training_set=E-HVC-146K, SPA-Compressor=w/o, scene_tokens (S)=64, event_tokens (E)=322026.01 | 13.94 | 1.46 | 31.32 | 68.34 | |
| Keye-VLevaluation_mode=Fine-tuned, training_set=E-HVC-146K2026.01 | 12.83 | 1.34 | 28.29 | 69.23 | |
| Qwen2.5-VL(7B)evaluation_mode=Fine-tuned2026.01 | 12.27 | 1.12 | 27.74 | 68.83 | |
| InternVL3(8B)evaluation_mode=Fine-tuned2026.01 | 11.98 | 1.04 | 27.38 | 67.75 | |
| Gemini 2.5 Proevaluation_mode=Zero-shot2026.01 | 4.3 | 1.41 | 22.35 | 66.58 | |
| Gemini 2.5 Flashevaluation_mode=Zero-shot2026.01 | 4.03 | 1.25 | 18.51 | 63.48 | |
| Qwen2.5-VL(72B)evaluation_mode=Zero-shot2026.01 | 3.31 | 1.24 | 22.21 | 67.2 | |
| Qwen2.5-VL(32B)evaluation_mode=Zero-shot2026.01 | 3.22 | 1.21 | 20.26 | 67.11 | |
| Qwen2.5-VL(7B)evaluation_mode=Zero-shot2026.01 | 2.91 | 0.59 | 13.29 | 65.82 | |
| GPT-4oevaluation_mode=Zero-shot2026.01 | 2.38 | 1.07 | 21.29 | 67.73 |