Text-to-Video temporal grounding on ChronusAV
1.16BLEU-4ChronusOmni
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ChronusOmniModel Size=7B2025.12 | 1.16 | 3.37 | 2.12 | 5.07 | |
| Qwen3-OmniModel Size=30B-A3B2025.12 | 0.37 | 2.13 | 1.62 | 0.92 | |
| LongVALE-LLMModel Size=7B2025.12 | 0.35 | 2.03 | 0.99 | 1.38 | |
| Qwen2.5-OmniModel Size=7B2025.12 | 0.35 | 1.89 | 1.02 | 0.79 | |
| OlaModel Size=7B2025.12 | 0.26 | 1.66 | 0.99 | 0.47 | |
| ARC-Hunyuan-VideoModel Size=7B2025.12 | 0.24 | 1.54 | 1.11 | 0.46 | |
| VideoLLaMAModel Size=7B2025.12 | 0.1 | 1.5 | 0.93 | 0.03 | |
| AVicunaModel Size=7B2025.12 | 0.04 | 1.03 | 0.32 | 0.12 |