search

HarryVolek / PyTorch_Speaker_Verification

PyTorch implementation of "Generalized End-to-End Loss for Speaker Verification" by Wan, Li et al.
BSD 3-Clause "New" or "Revised" License
575 stars 166 forks source link

The format of label is different between your code(dvector_create.py) and the example data in uis-rnn. #11

Closed wuqiangch closed 5 years ago

wuqiangch commented 5 years ago

@HarryVolek The label in example data in uis-rnn starts with the name of audio, but in your code the label has no the name of audio. Does it matter?

HarryVolek commented 5 years ago

Not quite sure what you mean, the labels just need to be unique for each speaker. I export a numpy string array of labels.