Audio Question Answering on MMAU mini v05.15.25 (test)
76.8AccuracyAQA-TTRL
Evaluation Results
| Method | Links | |
|---|---|---|
| AQA-TTRLBackbone=Qwen2.5-Omni 7B, Protocol=Self-Adaptation2025.10 | 76.8 | |
| SFTBackbone=Qwen2.5-Omni 7B, Protocol=Supervised Fine-Tuning2025.10 | 73.9 | |
| DIMVBackbone=Qwen2.5-Omni 7B, Protocol=Direct Inference with Majority Voting2025.10 | 73.3 | |
| DIBackbone=Qwen2.5-Omni 7B, Protocol=Direct Inference2025.10 | 72.4 | |
| AQA-TTRLBackbone=Qwen2.5-Omni 3B, Protocol=Self-Adaptation2025.10 | 72.3 | |
| SFTBackbone=Qwen2.5-Omni 3B, Protocol=Supervised Fine-Tuning2025.10 | 65.2 | |
| DIMVBackbone=Qwen2.5-Omni 3B, Protocol=Direct Inference with Majority Voting2025.10 | 63.9 | |
| DIBackbone=Qwen2.5-Omni 3B, Protocol=Direct Inference2025.10 | 61.5 |