Text-Audio Alignment on Baton-Pair three-sound-event
69.1Pair AccuracyAQAScore
Evaluation Results
| Method | Links | |
|---|---|---|
| AQAScoreBackbone=Qwen2.5-Omni-3B2026.01 | 69.1 | |
| AQAScoreBackbone=Qwen2.5-Omni-7B2026.01 | 67.8 | |
| AQAScoreBackbone=AF32026.01 | 63.2 | |
| AQAScoreBackbone=AF3-Think2026.01 | 62 | |
| CLAPScoreVariant=630k-best2026.01 | 61.9 | |
| AQAScoreBackbone=AF3-Chat2026.01 | 60.2 | |
| CLAPScoreTraining Data=Music+AS2026.01 | 60 | |
| CLAPScoreTraining Data=Music+Speech+AS2026.01 | 58 | |
| CLAPScoreTraining Data=630k-AS2026.01 | 57.5 | |
| CLAPScoreTraining Data=Music+Speech2026.01 | 56.8 | |
| PromptingBackbone=Qwen2.5-Omni-7B2026.01 | 54.3 | |
| PromptingBackbone=AF32026.01 | 50.7 | |
| PromptingBackbone=AF3-Chat2026.01 | 50.5 | |
| PromptingBackbone=Qwen2.5-Omni-3B2026.01 | 50.1 | |
| PromptingBackbone=AF3-Think2026.01 | 50 | |
| CascadeBackbone=AF3-Chat2026.01 | 42 | |
| CascadeBackbone=Qwen2.5-Omni-7B2026.01 | 38.2 | |
| CascadeBackbone=AF32026.01 | 38 | |
| CascadeBackbone=Qwen2.5-Omni-3B2026.01 | 35.6 | |
| CascadeBackbone=AF3-Think2026.01 | 22.6 |