Automatic Speech Recognition on TED-LIUM 3
3.7WEROffline Whisper
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Offline WhisperModel=large-v2, Chunk Size=-2025.08 | 3.7 | — | |
| Offline WhisperModel=small, Chunk Size=-2025.08 | 4.3 | — | |
| Offline WhisperModel=base, Chunk Size=-2025.08 | 5 | — | |
| Ufal-WhisperModel=small, Chunk Size=3002025.08 | 6.66 | 8.55 | |
| Simul-WhisperModel=small, Chunk Size=3002025.08 | 6.69 | 8.16 | |
| Simul-WhisperModel=small, Chunk Size=2002025.08 | 7.06 | 8.84 | |
| WhisperRTModel=large-v2, Chunk Size=3002025.08 | 7.22 | 8.7 | |
| WhisperRTModel=large-v2, Chunk Size=2002025.08 | 7.67 | 9.07 | |
| WhisperRTModel=large-v2, Chunk Size=1002025.08 | 7.71 | 9.8 | |
| WhisperRTModel=small, Chunk Size=3002025.08 | 7.81 | 9.84 | |
| Simul-WhisperModel=large-v2, Chunk Size=3002025.08 | 7.89 | 8.05 | |
| Simul-WhisperModel=large-v2, Chunk Size=2002025.08 | 8.08 | 8.31 | |
| Ufal-WhisperModel=base, Chunk Size=3002025.08 | 8.14 | 9.67 | |
| WhisperRTModel=small, Chunk Size=2002025.08 | 8.5 | 10.35 | |
| Ufal-WhisperModel=small, Chunk Size=2002025.08 | 8.52 | 10.22 | |
| Ufal-WhisperModel=large-v2, Chunk Size=3002025.08 | 8.82 | 8.95 | |
| WhisperRTModel=small, Chunk Size=1002025.08 | 9.69 | 12.24 | |
| Simul-WhisperModel=large-v2, Chunk Size=1002025.08 | 9.7 | 9.56 | |
| Simul-WhisperModel=small, Chunk Size=1002025.08 | 10.35 | 10.79 | |
| WhisperRTModel=large-v2, Chunk Size=402025.08 | 10.38 | 11.41 | |
| WhisperRTModel=base, Chunk Size=3002025.08 | 10.46 | 12.72 | |
| Ufal-WhisperModel=base, Chunk Size=2002025.08 | 10.63 | 12.38 | |
| WhisperRTModel=base, Chunk Size=2002025.08 | 10.74 | 13.33 | |
| Simul-WhisperModel=base, Chunk Size=3002025.08 | 11.4 | 12.46 | |
| Simul-WhisperModel=large-v2, Chunk Size=402025.08 | 11.64 | 12.75 | |
| WhisperRTModel=small, Chunk Size=402025.08 | 11.91 | 14.88 | |
| Simul-WhisperModel=base, Chunk Size=2002025.08 | 12.02 | 12.46 | |
| WhisperRTModel=base, Chunk Size=1002025.08 | 12.49 | 15.42 | |
| TASUTrain Data (Text)=Libri+Slide2026.04 | 13.23 | — | |
| Ufal-WhisperModel=large-v2, Chunk Size=2002025.08 | 13.31 | 13.65 | |
| TASU2Train Data (Text)=Libri+Slide, Conditioning=WER-binned2026.04 | 13.93 | — | |
| TASU2Train Data (Text)=Libri, Conditioning=None2026.04 | 14.02 | — | |
| Simul-WhisperModel=small, Chunk Size=402025.08 | 14.18 | 14.53 | |
| WhisperRTModel=base, Chunk Size=402025.08 | 14.4 | 17.56 | |
| TASU2Train Data (Text)=Libri, Conditioning=WER-binned2026.04 | 14.49 | — | |
| SLAM-CTCTrain Data (Audio)=Libri2026.04 | 14.61 | — | |
| Simul-WhisperModel=base, Chunk Size=1002025.08 | 14.92 | 15.28 | |
| Ufal-WhisperModel=small, Chunk Size=1002025.08 | 15.21 | 17.11 | |
| Ufal-WhisperModel=base, Chunk Size=1002025.08 | 17.81 | 19.62 | |
| TASUTrain Data (Text)=Libri2026.04 | 19.36 | — | |
| Simul-WhisperModel=base, Chunk Size=402025.08 | 24.57 | 23.76 | |
| Ufal-WhisperModel=large-v2, Chunk Size=1002025.08 | 30.51 | 30.25 | |
| Ufal-WhisperModel=small, Chunk Size=402025.08 | 34.11 | 37.08 | |
| Ufal-WhisperModel=base, Chunk Size=402025.08 | 36.4 | 37.77 | |
| Ufal-WhisperModel=large-v2, Chunk Size=402025.08 | 81.88 | 71.05 |