Multilingual Automatic Speech Recognition on CommonVoice
4.42WER (de)Qwen2.5-Omni + SICL-AT1
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen2.5-Omni + SICL-AT1few-shot=true2026.01 | 4.42 | 5.12 | 6.39 | |
| Qwen2.5-Omni + SICL-AT3few-shot=true2026.01 | 4.58 | 4.73 | 6.57 | |
| Qwen2.5-Omni + SICL-AT2few-shot=true2026.01 | 4.75 | 4.71 | 6.51 | |
| Qwen2.5-Omni + Vanilla SICLfew-shot=true2026.01 | 5.07 | 5.39 | 7.5 | |
| Qwen2.5-Omnifew-shot=false2026.01 | 5.75 | 5.15 | 7.8 | |
| Qwen2.5-Omni (Fine-tuned on CV-en) + Vanilla SICLfew-shot=true2026.01 | 8.16 | 8.29 | 10.53 | |
| Qwen2.5-Omni (Fine-tuned on RSR)few-shot=false2026.01 | 8.43 | 7.84 | 11.42 | |
| Qwen2.5-Omni (Fine-tuned on CV-en)few-shot=false2026.01 | 8.64 | 7.37 | 10.58 | |
| MiMo-Audio + SICL-AT2few-shot=true2026.01 | 30.49 | 6.51 | 39.47 | |
| MiMo-Audio + SICL-AT3few-shot=true2026.01 | 31.22 | 6.62 | 40.46 | |
| MiMo-Audio + SICL-AT1few-shot=true2026.01 | 34.11 | 6.59 | 45.5 | |
| MiMo-Audio + Vanilla SICLfew-shot=true2026.01 | 37.46 | 11.05 | 45.88 | |
| MiMo-Audiofew-shot=false2026.01 | 69.74 | 14.43 | 90.68 |