Automatic Speech Recognition on Spgispeech (test)
2.17WERSpeech-Hands
Evaluation Results
| Method | Links | |
|---|---|---|
| Speech-HandsFramework=Qwen2.5-Omni, External Reference=Canary-1b-v22026.01 | 2.17 | |
| Speech-HandsFramework=Qwen2.5-Omni, External Reference=Parakeet-tdt-0.6b-v32026.01 | 2.26 | |
| ReHear2026.02 | 2.36 | |
| ReHear + Rule2026.02 | 2.37 | |
| ReHear + Model2026.02 | 2.39 | |
| ISL2026.02 | 2.41 | |
| IPL + Rule2026.02 | 2.57 | |
| IPL2026.02 | 2.83 | |
| Speech-HandsFramework=Qwen2.5-Omni, External Reference=Whisper-v2-large2026.01 | 3.01 | |
| Canary-1b-v2Model Type=ASR model2026.01 | 3.08 | |
| Phi-4-MMModel Type=Omni-LLM2026.01 | 3.13 | |
| Parakeet-tdt-0.6b-v3Model Type=ASR model2026.01 | 3.16 | |
| Gemini-2-FlashModel Type=Omni-LLM2026.01 | 3.82 | |
| Whisper-v2-largeModel Type=ASR model2026.01 | 3.94 | |
| GER (Cascaded)Framework=Qwen2.5-Omni, External Reference=Whisper-v2-large2026.01 | 3.94 | |
| GER (Cascaded)Framework=Qwen2.5-Omni, External Reference=Parakeet-tdt-0.6b-v32026.01 | 3.98 | |
| GER (Cascaded)Framework=Qwen2.5-Omni, External Reference=Canary-1b-v22026.01 | 4.02 | |
| Qwen2.5_omniModel Type=Omni-LLM2026.01 | 4.58 | |
| GPT-4o-voiceModel Type=Omni-LLM2026.01 | 5.66 |