Speech Recognition on AISHELL-8-RealScene Outdoor (test)
37.47CERM2S-AVSR
Evaluation Results
| Method | Links | |
|---|---|---|
| M2S-AVSRSystem Type=AVSR, Modality=Audio-Visual2026.06 | 37.47 | |
| FireRed-ASRSystem Type=ASR, LLM-based=true2026.06 | 38.69 | |
| Qwen-ASRSystem Type=ASR, LLM-based=true2026.06 | 38.76 | |
| MMS-LLaMASystem Type=AVSR, LLM-based=true2026.06 | 39.16 | |
| M2S-AVSR (Audio-only)System Type=AVSR, Modality=Audio-only2026.06 | 40.56 | |
| LLaMA-AVSRSystem Type=AVSR, LLM-based=true2026.06 | 40.82 | |
| Whisper-FlamingoSystem Type=AVSR2026.06 | 41.64 | |
| Whisper-finetunedSystem Type=ASR2026.06 | 42.01 | |
| Fun-ASR-NanoSystem Type=ASR, LLM-based=true2026.06 | 49.94 |