List your company
ResearchFormerly WizWand

Automatic Speech Recognition on Mandarin Stuttered Speech (Severe)

5.17CERQwen3-ASR-0.6B
4.038811.674419.3126.9456Jun 19, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Qwen3-ASR-0.6BFine-tuned Dataset=AS-70 + Synthetic2026.065.17
Wav2Vec2.0Fine-tuned Dataset=AS-70 + Synthetic2026.068.26
Whisper-large-v3Fine-tuned Dataset=AS-70 + Synthetic2026.069.27
Qwen3-ASR-0.6BFine-tuned Dataset=AS-702026.069.36
Wav2Vec2.0Fine-tuned Dataset=AS-702026.0611.09
Whisper-large-v3Fine-tuned Dataset=AS-702026.0617.36
Qwen3-ASR-0.6BFine-tuned Dataset=None2026.0619.74
Wav2Vec2.0Fine-tuned Dataset=None2026.0630.7
Whisper-large-v3Fine-tuned Dataset=None2026.0633.45