Automatic Speech Recognition on AISHELL-2 mic
2.5CERQwen2.5-Omni
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5-OmniModel Size=7B2025.08 | 2.5 | 2.5 | |
| Kimi-Audio-InstructModel Size=7B2025.08 | 2.7 | 2.7 | |
| OSUM2026.03 | 2.81 | 2.81 | |
| Qwen2-Audio2026.03 | 3 | 3 | |
| OSUM-Pangu2026.03 | 3.01 | 3.01 | |
| MiDashengLMModel Size=7B2025.08 | 3.2 | 3.2 | |
| UMA + self-conditionModel Type=NAR, Self-conditioning=true, RTF=0.098, Params(M)=110.42023.09 | 5.9 | — | |
| UMAModel Type=NAR, Self-conditioning=false, RTF=0.085, Params(M)=105.42023.09 | 6 | — | |
| Hybrid (Conformer) + beam searchModel Type=AR, Beam search=true, RTF=0.954, Params(M)=116.42023.09 | 6.1 | — | |
| CIF + SANModel Type=NAR2023.09 | 6.3 | — | |
| Hybrid (Conformer)Model Type=AR, Beam search=false, RTF=0.205, Params(M)=116.42023.09 | 6.8 | — | |
| LASO-largeModel Type=NAR, Params(M)=80.02023.09 | 7.4 | — |