taylorlu / ghostvlad-speaker

An tensorflow implementation of ghostvlad for speaker recognition

ghostvlad-speaker

An tensorflow implementation of ghostvlad, pretrained model can be downloaded in ghostvlad-speaker.tar

prerequisit

1. pydub, also need ffmpeg to be installed.
2. tensorflow
3. librosa

Steps:

1. generate speaker labels from dataset voxceleb2

python voxcele2json.py

2. Training

python train.py
Please change the args_params in main

3. predict

python predict.py