Automated Audio Captioning on MACS
64.9AAC ScoreKimi-Audio
Evaluation Results
| Method | Links | |
|---|---|---|
| Kimi-AudioSize=7B2025.09 | 64.9 | |
| SLAM-OmniSize=0.5B2025.09 | 50.46 | |
| TtTSize=3B2025.09 | 48.87 | |
| Qwen2.5-3B (AR)Size=3B2025.09 | 48.64 | |
| Pretrain+TtTSize=3B2025.09 | 42.86 | |
| Qwen2.5-3B (NAR)Size=3B2025.09 | 27.4 | |
| GLM-4-VoiceSize=9B2025.09 | 12.67 | |
| MoshiSize=7B2025.09 | 12.01 | |
| VITA-AudioSize=7B2025.09 | 7.94 | |
| LLaMA-OmniSize=8B2025.09 | 4.56 | |
| Mini-OmniSize=0.5B2025.09 | 4.45 | |
| SpeechGPTSize=7B2025.09 | 3.95 |