Speaker Recognition Metrics Datasets Analysis Output 'xvector.txt' format Get-Content "data/xvector.txt" | select -First 50 > "data/xvector-sample.txt" Metrics Equal Error Rate (EER) minDCF (Minimum Detection Cost Function (DCF) Calibrated Log-Likelihood Ratio (Cllr) References https://github.com/clovaai/voxceleb_trainer https://github.com/bsxfan/PYLLR