Child's Automatic Speech Recognition on RSR
12.19WERTICL+
Evaluation Results
| Method | Links | |
|---|---|---|
| TICL+Model=Phi-4-MM, k=32025.12 | 12.19 | |
| TICL+Model=Phi-4-MM, k=22025.12 | 12.26 | |
| TICL+Model=Phi-4-MM, k=42025.12 | 12.75 | |
| TICL+Model=Phi-4-MM, k=12025.12 | 12.89 | |
| MiMo-Audio + SICL-AT1few-shot=true, version=AT12026.01 | 16.59 | |
| MiMo-Audio + Vanilla SICLfew-shot=true2026.01 | 16.84 | |
| MiMo-Audio + SICL-AT2few-shot=true, version=AT22026.01 | 16.89 | |
| MiMo-Audio + SICL-AT3few-shot=true, version=AT32026.01 | 16.95 | |
| TICLModel=Phi-4-MM, k=32025.12 | 18.9 | |
| TICLModel=Phi-4-MM, k=12025.12 | 18.92 | |
| TICLModel=Phi-4-MM, k=22025.12 | 18.92 | |
| TICLModel=Phi-4-MM, k=42025.12 | 19.54 | |
| Zero-ShotModel=Phi-4-MM, k=02025.12 | 20.06 | |
| Qwen2.5-Omni + SICL-AT1few-shot=true, version=AT12026.01 | 20.96 | |
| Qwen2.5-Omni + SICL-AT2few-shot=true, version=AT22026.01 | 21.95 | |
| Qwen2.5-Omni + SICL-AT3few-shot=true, version=AT32026.01 | 22.16 | |
| Qwen2.5-Omni (Fine-tuned on CV-en) + Vanilla SICLfew-shot=true, fine-tuning=CommonVoice-en2026.01 | 23.62 | |
| Qwen2.5-Omni + Vanilla SICLfew-shot=true2026.01 | 27.86 | |
| Qwen2.5-Omni (Fine-tuned on CV-en)few-shot=false, fine-tuning=CommonVoice-en2026.01 | 30.61 | |
| Qwen2.5-Omni (Fine-tuned on RSR)few-shot=false, fine-tuning=RSR2026.01 | 31.09 | |
| MiMo-Audiofew-shot=false2026.01 | 31.39 | |
| Qwen2.5-Omnifew-shot=false2026.01 | 35.65 |