Automatic Speech Recognition on Mandarin Stuttered Speech (Mild)
3.57CERQwen3-ASR-0.6B
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-ASR-0.6BFine-tuned Dataset=AS-70 + Synthetic2026.06 | 3.57 | |
| Whisper-large-v3Fine-tuned Dataset=AS-70 + Synthetic2026.06 | 4.49 | |
| Qwen3-ASR-0.6BFine-tuned Dataset=AS-702026.06 | 4.88 | |
| Whisper-large-v3Fine-tuned Dataset=AS-702026.06 | 4.97 | |
| Wav2Vec2.0Fine-tuned Dataset=AS-70 + Synthetic2026.06 | 6.78 | |
| Qwen3-ASR-0.6BFine-tuned Dataset=None2026.06 | 7.01 | |
| Wav2Vec2.0Fine-tuned Dataset=AS-702026.06 | 9.25 | |
| Whisper-large-v3Fine-tuned Dataset=None2026.06 | 13.26 | |
| Wav2Vec2.0Fine-tuned Dataset=None2026.06 | 24.04 |