Speech-to-text translation on MuST-C
37Average BLEUVIRBA
Evaluation Results
| Method | Links | |
|---|---|---|
| VIRBABackbone=Qwen2.5-Omni2025.08 | 37 | |
| Qwen2.5-Omni2025.08 | 36 | |
| Kimi-Audio2025.08 | 35.8 | |
| Step-Audio-R12025.08 | 35 | |
| AF32025.08 | 34.5 | |
| VIRBABackbone=Qwen2-Audio2025.08 | 34.4 | |
| Group-DPO2025.08 | 33.5 | |
| SeamlessM4T-L V22025.08 | 31.9 | |
| Qwen2-Audiovariant=base2025.08 | 29.1 | |
| Whisper-large V22025.08 | 25.4 |