Text-Audio Alignment on Baton two-sound-event
0.69AUCAQAScore
Evaluation Results
| Method | Links | |
|---|---|---|
| AQAScoreBackbone=Qwen2.5-Omni-7B2026.01 | 0.69 | |
| AQAScoreBackbone=Qwen2.5-Omni-3B2026.01 | 0.68 | |
| AQAScoreBackbone=AF3-Think2026.01 | 0.66 | |
| AQAScoreBackbone=AF32026.01 | 0.64 | |
| CascadeBackbone=Qwen2.5-Omni-7B2026.01 | 0.59 | |
| CLAPScoreVariant=630k-best2026.01 | 0.58 | |
| AQAScoreBackbone=AF3-Chat2026.01 | 0.57 | |
| CLAPScoreTraining Data=Music+AS2026.01 | 0.57 | |
| CLAPScoreTraining Data=Music+Speech+AS2026.01 | 0.56 | |
| CLAPScoreTraining Data=Music+Speech2026.01 | 0.55 | |
| CLAPScoreTraining Data=630k-AS2026.01 | 0.53 | |
| CascadeBackbone=Qwen2.5-Omni-3B2026.01 | 0.52 | |
| PromptingBackbone=Qwen2.5-Omni-3B2026.01 | 0.5 | |
| PromptingBackbone=Qwen2.5-Omni-7B2026.01 | 0.5 | |
| PromptingBackbone=AF32026.01 | 0.5 | |
| PromptingBackbone=AF3-Think2026.01 | 0.5 | |
| PromptingBackbone=AF3-Chat2026.01 | 0.5 | |
| CascadeBackbone=AF32026.01 | 0.5 | |
| CascadeBackbone=AF3-Think2026.01 | 0.5 | |
| CascadeBackbone=AF3-Chat2026.01 | 0.5 |