Audio Deepfake Detection on ADD All-type v1 (full)
99.75Speech AccuracyFT-GRPO
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FT-GRPOExplainable=true2026.01 | 99.75 | 75.05 | 84.26 | 90.44 | 87.38 | |
| W2V2-AASISTExplainable=false2026.01 | 99.58 | 37.17 | 59.08 | 58.23 | 63.52 | |
| ALLM4ADDExplainable=false2026.01 | 99.43 | 72 | 80.85 | 91.09 | 85.84 | |
| WPT-W2V2-AASISTExplainable=false2026.01 | 99.06 | 46.37 | 38.56 | 92.87 | 69.22 |