Phone Recognition on PRiSM Accented English Datasets
8.3PFER (Timing)HuPER-Recognizer
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| HuPER-RecognizerModel Category=English-centric PR Models, Dataset included during training=PR-tmt, PR-arc, PR-saa2026.03 | 8.3 | 12.5 | 11.6 | 10.8 | |
| KoelLabs-XLSRModel Category=English-centric PR Models, Dataset included during training=PR-tmt, PR-arc, PR-saa2026.03 | 9.6 | 7.8 | 7.8 | 8.4 | |
| ZIPA-CTCModel Category=Multilingual PR Models2026.03 | 13.1 | 9.7 | 9 | 10.6 | |
| ZIPA-CTC-NSModel Category=Multilingual PR Models2026.03 | 13.1 | 9.7 | 8.9 | 10.6 | |
| POWSM-CTCModel Category=Multilingual PR Models2026.03 | 13.1 | 10.3 | 10 | 11.1 | |
| W2V2P-LV60Model Category=Multilingual PR Models, Dataset included during training=PR-tmt2026.03 | 13.2 | 10.9 | 9.4 | 11.2 | |
| PhoneticXEUSModel Category=Multilingual PR Models2026.03 | 13.3 | 9.9 | 8.5 | 10.6 | |
| W2V2P-XLSR53Model Category=Multilingual PR Models2026.03 | 13.5 | 9.9 | 9 | 10.8 | |
| POWSMModel Category=Multilingual PR Models2026.03 | 13.7 | 11.3 | 27.6 | 17.5 | |
| Qwen3-Omni-InstructModel Category=Large Audio Language Models, Uncertainty about training data=true2026.03 | 15.1 | 11.9 | 9.1 | 12 | |
| Gemini 2.5 FlashModel Category=Large Audio Language Models, Uncertainty about training data=true2026.03 | 15.2 | 12.7 | 13.2 | 13.7 | |
| MultiIPAModel Category=Multilingual PR Models2026.03 | 16.3 | 15.5 | 13.8 | 15.2 |