clovaai / voxceleb_trainer

In defence of metric learning for speaker recognition

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

How to calculate the test accuracy?

dairuining opened this issue · comments

Hello! I want to ask how to calculate the test accuracy?

EER( equal error rate) is used as metric in speaker recognition according to their paper, In defence of metric learning for speaker recognition . The code is defined as tuneThresholdfromScore() in tuneThreshold.py.