ge2e-speaker-verification
ge2e-speaker-verification copied to clipboard
funcwj
Pytorch implementation of "Generalized End-to-End Loss for Speaker Verification"
Speaker Verification with GE2E Loss
Pytorch implement of "Generalized End-to-End Loss for Speaker Verification"
Data Processing
- Vad (recommend py-webrtcvad)
- Log mel-spectrogram features (recommend librosa)
- Prepare data as
data/{train,dev}/{feats.scp,spk2utt}
Usage
see train.sh and compute_dvector.py
Reference
Wan L, Wang Q, Papir A, et al. Generalized end-to-end loss for speaker verification[C]//2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2018: 4879-4883.
funcwj