Free Q&A on LibriTTS + DEMAND mixtures Background
64.9ROUGE-LQwen2-Audio
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen2-AudioSystem Setup=Upper Bound: Oracle Speaker + Auditory LLM2025.02 | 64.9 | 67.3 | 94 | |
| AAD-LLMSystem Setup=Intention-Informed AAD-LLM2025.02 | 63.1 | 65 | 93.7 | |
| Qwen2-AudioSystem Setup=Proposed Baseline: Extracted Speaker + Auditory LLM2025.02 | 51.2 | 52.7 | 91.6 | |
| Qwen2-AudioSystem Setup=Lower Bound: Random Speaker + Auditory LLM2025.02 | 46.9 | 48.3 | 90.7 | |
| SALMONNSystem Setup=Auditory LLM without Intention2025.02 | 43.6 | 45.5 | 89.9 | |
| Qwen2-AudioSystem Setup=Auditory LLM without Intention2025.02 | 40.2 | 41.3 | 89.4 | |
| WavLLMSystem Setup=Auditory LLM without Intention2025.02 | 39.1 | 37.4 | 89.1 | |
| LTU-ASSystem Setup=Auditory LLM without Intention2025.02 | 32.6 | 36.9 | 88.4 | |
| Qwen-AudioSystem Setup=Auditory LLM without Intention2025.02 | 30.7 | 35.5 | 87.5 | |
| GAMASystem Setup=Auditory LLM without Intention2025.02 | 23.4 | 27.9 | 86.7 |