List your company

Text-Audio Alignment on Baton three-sound-event

0.65AUCAQAScore
0.4940.53450.5750.6155Jan 21, 2026

Evaluation results

MethodLinks
AQAScoreBackbone=Qwen2.5-Omni-7B2026.010.65
CLAPScoreTraining Data=Music+Speech2026.010.63
CLAPScoreTraining Data=630k-AS2026.010.62
CLAPScoreVariant=630k-best2026.010.62
AQAScoreBackbone=Qwen2.5-Omni-3B2026.010.61
AQAScoreBackbone=AF3-Think2026.010.61
AQAScoreBackbone=AF32026.010.6
CLAPScoreTraining Data=Music+AS2026.010.56
CascadeBackbone=Qwen2.5-Omni-3B2026.010.55
AQAScoreBackbone=AF3-Chat2026.010.54
CLAPScoreTraining Data=Music+Speech+AS2026.010.54
CascadeBackbone=Qwen2.5-Omni-7B2026.010.53
CascadeBackbone=AF3-Chat2026.010.51
PromptingBackbone=Qwen2.5-Omni-3B2026.010.5
PromptingBackbone=Qwen2.5-Omni-7B2026.010.5
PromptingBackbone=AF32026.010.5
PromptingBackbone=AF3-Think2026.010.5
PromptingBackbone=AF3-Chat2026.010.5
CascadeBackbone=AF32026.010.5
CascadeBackbone=AF3-Think2026.010.5