Automatic Speech Recognition on MGB2 (test)
12WERAura-STT-LF
Evaluation Results
| Method | Links | |
|---|---|---|
| Aura-STT-LFParams=479M2026.03 | 12 | |
| Aura-STT*Params=446M, configuration=VAD + Aura-STT (short-form), beam-size=1 (greedy search)2026.03 | 12.62 | |
| Whisper-large-v3Params=1.55B2026.03 | 15.98 | |
| Nvidia-Conformer-CTCParams=121M2026.03 | 17.81 | |
| OWSM-CTC-v4Params=1.0B2026.03 | 71.56 |