Text-to-Audio temporal grounding on ChronusAV
7.4BLEU-4PadCaptioner
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| PadCaptionerModel Parameters=3B2026.07 | 7.4 | 7.3 | 4.8 | 35.1 | |
| ChronusOmniModel Parameters=7B2026.07 | 6.8 | 6.9 | 4.5 | 34.3 | |
| ChronusOmniModel Size=7B2025.12 | 6.78 | 6.86 | 4.5 | 34.3 | |
| Qwen3-OmniModel Size=30B-A3B2025.12 | 0.94 | 2.26 | 1.2 | 2.18 | |
| Qwen3-OmniModel Parameters=30B-A3B2026.07 | 0.9 | 2.3 | 1.2 | 2.2 | |
| Qwen2.5-OmniModel Parameters=7B2026.07 | 0.7 | 1 | 0.6 | 0.3 | |
| Qwen2.5-OmniModel Size=7B2025.12 | 0.67 | 1.01 | 0.64 | 0.26 | |
| OlaModel Parameters=7B2026.07 | 0.2 | 0.8 | 0.3 | 0.3 | |
| LongVALE-LLMModel Parameters=7B2026.07 | 0.2 | 1.3 | 0.5 | 0.2 | |
| ARC-HY-VideoModel Parameters=7B2026.07 | 0.2 | 1.2 | 0.7 | 0.4 | |
| OlaModel Size=7B2025.12 | 0.18 | 0.78 | 0.31 | 0.26 | |
| ARC-Hunyuan-VideoModel Size=7B2025.12 | 0.18 | 1.17 | 0.65 | 0.37 | |
| LongVALE-LLMModel Size=7B2025.12 | 0.15 | 1.31 | 0.5 | 0.19 | |
| VideoLLaMAModel Parameters=7B2026.07 | 0.1 | 1.1 | 0.6 | 0.1 | |
| VideoLLaMAModel Size=7B2025.12 | 0.05 | 1.13 | 0.63 | 0.06 | |
| AVicunaModel Size=7B2025.12 | 0.01 | 0.51 | 0.13 | 0.02 | |
| AVicunaModel Parameters=7B2026.07 | 0 | 0.5 | 0.1 | 0 |