search

YChenL / DS-TDNN

Official implement of "Dual-stream Time-Delay Neural Network with Dynamic Global Filter for Speaker Verification" in PyTorch
https://arxiv.org/pdf/2303.11020v2.pdf
40 stars 7 forks source link
automatic-speaker-verification deep-learning nerual-network time-delay-neural-network

readme

Detials will coming soon. (Updating)

This repo includes implements of several SOTA front-end models for ASV

  1. ECAPA-TDNN (Un-official)
  2. SE-ResNet (Un-official)
  3. MFA-TDNN (Un-official)
  4. MFA-Conformer (Official)
  5. DS-TDNN (Ours)

and some popular loss functions for ASV

  1. AMSoftmax
  2. AAMSoftmax
  3. Sub-center Arcface
  4. ARMSoftmax

Overviwes of DS-TDNN

Architecture

Experiment Results