Automatic Speech Recognition on LibriSpeech LS-other
3.49WERQwen3-ASR-1.7B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-ASR-1.7BDecoding Architecture=AR, Params (B)=2.02025.08 | 3.49 | 20.12 | 5.62 | |
| Whisfusion + rerankDecoding Architecture=NAR, Params (B)=2.42025.08 | 3.53 | 143.27 | 6.07 | |
| Canary-1b-v2Decoding Architecture=AR, Params (B)=1.02025.08 | 3.68 | 47.87 | 6.4 | |
| WhisfusionDecoding Architecture=NAR, Params (B)=1.52025.08 | 3.78 | 173.45 | 6.55 | |
| Whisper-turboDecoding Architecture=AR, Params (B)=0.82025.08 | 4.05 | 143.99 | 8.86 | |
| Whisper-large-v3Decoding Architecture=AR, Params (B)=1.62025.08 | 4.31 | 35.33 | 7.11 | |
| Whisper-LLaDADecoding Architecture=NAR, Params (B)=8.72025.08 | 4.88 | 7.92 | 9.82 | |
| OWSM-CTC v3.1Decoding Architecture=NAR, Params (B)=1.02025.08 | 5.3 | 665.74 | 8.27 | |
| Drax (MBR, 8/16)Decoding Architecture=NAR, Params (B)=1.22025.08 | 5.34 | 33.16 | 8.34 | |
| MMS-allDecoding Architecture=NAR, Params (B)=1.02025.08 | 8.08 | 681.21 | 12.93 |