Speaker Verification on VoxCeleb2 O (test)
1.005EERE.-H. AM-Softmax
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| E.-H. AM-SoftmaxMargin=true, Scale=true2026.01 | 1.005 | 0.077 | |
| HAM-SoftmaxMargin=true, Scale=true2026.01 | 1.048 | 0.072 | |
| Real AM-SoftmaxMargin=true, Scale=true2026.01 | 1.085 | 0.083 | |
| AAM-SoftmaxMargin=true, Scale=true2026.01 | 1.249 | 0.092 | |
| AM-SoftmaxMargin=true, Scale=true2026.01 | 1.298 | 0.102 | |
| H-SoftmaxMargin=false, Scale=true2026.01 | 1.468 | 0.111 | |
| OR-Gatenoise_ratio (η)=10%2022.11 | 1.62 | — | |
| LNCL+Sub-AM + Cosinenoise_ratio (η)=0%, scoring_method=Cosine2022.11 | 1.63 | — | |
| LNCL+Sub-AM + Cosinenoise_ratio (η)=5%, scoring_method=Cosine2022.11 | 1.63 | — | |
| LNCL+Sub-AM + Cosinenoise_ratio (η)=10%, scoring_method=Cosine2022.11 | 1.64 | — | |
| OR-Gatenoise_ratio (η)=0%2022.11 | 1.64 | — | |
| OR-Gatenoise_ratio (η)=5%2022.11 | 1.65 | — | |
| OR-Gatenoise_ratio (η)=20%2022.11 | 1.67 | — | |
| Baselinenoise_ratio (η)=0%2022.11 | 1.69 | — | |
| LNCL+Sub-AM + Cosinenoise_ratio (η)=20%, scoring_method=Cosine2022.11 | 1.69 | — | |
| LNCL+Sub-AM + NL-PLDAnoise_ratio (η)=0%, scoring_method=NL-PLDA2022.11 | 1.7 | — | |
| LNCL+Sub-AM + PLDAnoise_ratio (η)=0%, scoring_method=PLDA2022.11 | 1.71 | — | |
| Baselinenoise_ratio (η)=5%2022.11 | 1.72 | — | |
| LNCL+Sub-AM + NL-PLDAnoise_ratio (η)=5%, scoring_method=NL-PLDA2022.11 | 1.72 | — | |
| OR-Gatenoise_ratio (η)=30%2022.11 | 1.72 | — | |
| LNCL+Sub-AM + NL-PLDAnoise_ratio (η)=10%, scoring_method=NL-PLDA2022.11 | 1.73 | — | |
| LNCL+Sub-AM + NL-PLDAnoise_ratio (η)=20%, scoring_method=NL-PLDA2022.11 | 1.75 | — | |
| LNCL+Sub-AM + NL-PLDAnoise_ratio (η)=30%, scoring_method=NL-PLDA2022.11 | 1.77 | — | |
| LNCL+Sub-AM + PLDAnoise_ratio (η)=5%, scoring_method=PLDA2022.11 | 1.78 | — | |
| LNCL+Sub-AM + Cosinenoise_ratio (η)=30%, scoring_method=Cosine2022.11 | 1.8 | — | |
| LNCL+Sub-AM + PLDAnoise_ratio (η)=10%, scoring_method=PLDA2022.11 | 1.81 | — | |
| Baselinenoise_ratio (η)=10%2022.11 | 1.9 | — | |
| OR-Gatenoise_ratio (η)=50%2022.11 | 1.97 | — | |
| LNCL+Sub-AM + PLDAnoise_ratio (η)=20%, scoring_method=PLDA2022.11 | 2.05 | — | |
| SoftmaxMargin=false, Scale=false2026.01 | 2.074 | 0.146 | |
| LNCL+Sub-AM + PLDAnoise_ratio (η)=30%, scoring_method=PLDA2022.11 | 2.15 | — | |
| LNCL+Sub-AM + NL-PLDAnoise_ratio (η)=50%, scoring_method=NL-PLDA2022.11 | 2.19 | — | |
| Baselinenoise_ratio (η)=20%2022.11 | 2.21 | — | |
| LNCL+Sub-AM + Cosinenoise_ratio (η)=50%, scoring_method=Cosine2022.11 | 2.25 | — | |
| LNCL+Sub-AM + PLDAnoise_ratio (η)=50%, scoring_method=PLDA2022.11 | 2.86 | — | |
| Baselinenoise_ratio (η)=30%2022.11 | 2.88 | — | |
| SoftmaxMargin=false, Scale=true2026.01 | 3.046 | 0.217 | |
| Baselinenoise_ratio (η)=50%2022.11 | 4.32 | — |