Automatic Speech Recognition on MLS FR (test)
3.8WERMDM-ASR
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MDM-ASRArchitecture Type=NAR2026.02 | 3.8 | — | — | |
| Canary-1b-flashArchitecture Type=AR2026.02 | 4.5 | — | — | |
| Whisper-large-v3Architecture Type=AR2026.02 | 4.8 | — | — | |
| Phi-4-multimodalArchitecture Type=AR2026.02 | 4.9 | — | — | |
| Voxtral-MiniArchitecture Type=AR2026.02 | 5.3 | — | — | |
| SeamlessM4T-large-v2Architecture Type=AR2026.02 | 5.4 | — | — | |
| XLSR-53Architecture Type=NAR2026.02 | 5.6 | — | — | |
| DraxArchitecture Type=NAR2026.02 | 7.1 | — | — | |
| OWSM-v3.1Architecture Type=AR2026.02 | 12.1 | — | — | |
| Ufal-WhisperModel Size=large-v2, Chunk Size=300 msec, Initial Chunk Size=600 msec2025.08 | 12.4 | 16.79 | 15.94 | |
| OWSM-CTCArchitecture Type=NAR2026.02 | 12.9 | — | — | |
| WhisperRTModel Size=large-v2, Chunk Size=300 msec, Initial Chunk Size=600 msec2025.08 | 14.23 | 14.14 | 17.24 | |
| Simul-WhisperModel Size=large-v2, Chunk Size=300 msec, Initial Chunk Size=600 msec2025.08 | 15.24 | 17.4 | 20.65 |