search

Janghyun1230 / Speaker_Verification

Tensorflow implementation of "Generalized End-to-End Loss for Speaker Verification"
MIT License
354 stars 102 forks source link

new speaker #13

Open Aurora11111 opened 5 years ago

Aurora11111 commented 5 years ago

hello，can this project be used to identify a new speaker? I mean a wav which belongs to none of the enrollment utterance.

Sadam1195 commented 5 years ago

As per my understanding, NO. This project does not correspond the use case of new (unknown/not trained) speaker.