Speaker Diarization on NIST-SRE 2000
Metric: DER(%) (higher is better)
Metric: DER(%) (higher is better)
| # | Model↕ | DER(%)▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | x-vector (PLDA + AHC) | 8.39 | No | TitaNet: Neural Model for speaker representation... | 2021-10-08 | Code |
| 2 | TitaNet-L (NME-SC) | 6.73 | No | TitaNet: Neural Model for speaker representation... | 2021-10-08 | Code |
| 3 | TitaNet-M (NME-SC) | 6.47 | No | TitaNet: Neural Model for speaker representation... | 2021-10-08 | Code |
| 4 | TitaNet-S (NME-SC) | 6.37 | No | TitaNet: Neural Model for speaker representation... | 2021-10-08 | Code |
| 5 | x-vector (MCGAN) | 5.73 | No | TitaNet: Neural Model for speaker representation... | 2021-10-08 | Code |