Pathological Speech Reconstruction on SJTU Parkinson Patient Speech Dataset benchmark leaderboard · SOTA2 Research

Pathological Speech Reconstruction on SJTU Parkinson Patient Speech Dataset

0.1592WERQwen2.5-Omni

Evaluation Results

Method	Links
Qwen2.5-Omni2026.06		0.1592	0.0701	69.92	91.41
Audio-Flamingo 32026.06		0.1615	0.0717	69.61	90.79
Audio-Flamingo 32026.06		0.1682	0.0783	69.38	91.07
Qwen2.5-Omni2026.06		0.1724	0.0799	68.56	90.87
Gemini-Flash-3.52026.06		0.1817	0.0925	58.78	90.23
Qwen2.5-Omni2026.06		0.209	0.1058	64	88.47
Colm-DSR2026.06		0.216	0.093	59.5	88.64
Align-SLM2026.06		0.2284	0.0957	61.57	90.55
Diff-DSR2026.06		0.2312	0.1161	59.86	88.35
Audio-Flamingo 32026.06		0.2426	0.0949	61.09	90.22