/Model-Compression-For-Speaker-Recognition

Distillation examples. Trying to make Speaker Recognition Faster through different Model Compression techniques

Primary LanguagePython

Model Compression For DNN-Based Speaker Recognition

  • Trying to make Speaker Recognition Faster through different Model Compression techniques
  • Author: Liang Chen
  • Framework: PyTorch
  • Dataset: VoxCeleb1, TIMIT, CommonVoice, CN_Celeb, VCC2016

Todo List

  • Finish Example Experiment
  • Knowledge Distillation of X-vector system
  • Model Pruning of X-vector system
  • to be continued