Multi-turn audio dialogue on AF-Dialogue-AudioSetSL (test)
1.622CIDErAudio Flamingo
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Audio FlamingoEvaluation Protocol=fine-tuned2024.02 | 1.622 | 0.237 | 0.473 | |
| LTUEvaluation Protocol=fine-tuned2024.02 | 0.823 | 0.153 | 0.403 | |
| LTUEvaluation Protocol=zero-shot2024.02 | 0.58 | 0.122 | 0.324 | |
| Qwen-AudioEvaluation Protocol=zero-shot2024.02 | 0.507 | 0.06 | 0.292 |