Partially Fake Speech Localization on Half-Truth (HT)
0.1EERHT-finetuned Baseline
Evaluation results
| Method | Links | |
|---|---|---|
| HT-finetuned BaselineEvaluation Protocol=In-domain finetuning, Resolution=20 ms2026.04 | 0.1 | |
| Alethia-BaseEvaluation Protocol=Zero-shot testing, Resolution=20 ms2026.04 | 9.2 | |
| Alethia-LargeEvaluation Protocol=Zero-shot testing, Resolution=20 ms2026.04 | 10 | |
| W2V-1BEvaluation Protocol=Zero-shot testing, Resolution=20 ms2026.04 | 14.2 | |
| WavLM-LargeEvaluation Protocol=Zero-shot testing, Resolution=20 ms2026.04 | 18.1 | |
| W2V-300MEvaluation Protocol=Zero-shot testing, Resolution=20 ms2026.04 | 31.5 | |
| HuBERT-LargeEvaluation Protocol=Zero-shot testing, Resolution=20 ms2026.04 | 32.9 | |
| PS-finetuned BaselineEvaluation Protocol=In-domain finetuning, Resolution=20 ms2026.04 | 46.4 |