Query on comparison to the x-vector extractors developed by kaldi
#4
by uyiosa - opened
How does this compare to the x-vector extractors developed by kaldi? I'm looking for open source embedding extraction models that are best for voice specifically. I want good vector representations for speaker recognition. I've tried pyannote inference for embedding extraction which was not good enough for me as I kept getting alot of false positives and negatives with cosine similarity. I'm now looking at this model as well as Kaldi as they were recommended as they both extract x-vectors specifically.