my-yy / Speaker-Recognition

PyTorch implementation using high-level framework Catalyst of the paper Utterance-level Aggregation For Speaker Recognition In The Wild

README

Weights coming soon

Next step

• Add this model as a uis-rnn backend

This repo contains a Pytorch implementation of the paper,
Utterance-level Aggregation For Speaker Recognition In The Wild (Xie et al., ICASSP 2019) (Oral).

**New challenge on speaker recognition: The VoxCeleb Speaker Recognition Challenge (VoxSRC).

Dependencies

• PyTorch 1.3.1
• Catalyst 19.11.6
• Apex 0.1

Data

The dataset used for the experiments are

• Voxceleb1, Voxceleb2

Accuracy metrics