Can this be used for speaker identification/authentication/detection?
#1
by
emraza110
- opened
What are embeddings used for?
Hi who are you?
Yes, maybe you also want to check the wespeaker repo for direct application on speaker recognition tasks, the model here is the resnet34_LM listed in https://github.com/wenet-e2e/wespeaker/blob/master/docs/pretrained.md
Is this model different from pyannote/embedding. I used both of them, they just give the same results.
They are the same model
I can correctly get 512 dim feature from pyannote/embedding, but for this repo pyannote/wespeaker-voxceleb-resnet34-LM , it generates 256 dim feature with [256 * NaN]. Any instruction? many thanks.....