speaker-embeddings

Official repository for RawNet, RawNet2, and RawNet3

Deep speaker embeddings in PyTorch, including x-vectors. Code used in this work: https://arxiv.org/abs/2007.16196

Implementation of Neural Voice Cloning with Few Samples Research Paper by Baidu

Speaker identification/verification models for Machine Learning for Computer Vision class at UNIBO

[ICASSP'23] Online speaker clustering

The ITMO Speaker Recognition Online Course is a free educational course covering classical and modern methods of speaker recognition. The course provides video lectures, lecture materials, and practical laboratory assignments (Python/PyTorch).

Angular triplet center loss implementation in Pytorch.

This repository contain the code of the main part of my master thesis degree at Politecnico di Torino in Data science & Engineering

Speaker diarization on VoxData — benchmarked state-of-the-art speaker embedding models against each other, built a custom diarization model, and compared clustering strategies.

ReDimNet2 inference w/ ONNX

I. Thoidis, C. Gaultier, and T. Goehring, "Perceptual Analysis of Speaker Embeddings for Voice Discrimination between Machine And Human Listening," ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Rhodes Island, Greece, 2023, pp. 1-5

Persistent speaker identity across audio files — a local voice registry that resolves diarization clusters to named speakers

Multi-resolution speaker embeddings with ReDimNet architecture and Matryoshka Representation Learning (MRL)

Deterministic checks for speaker-verification studies: detect F0-tracking contamination that can invert your statistics. Zero dependencies, no audio required.

Research about how to design a robot sound from voice conversion and speaker embeddings

Generation of Novel Human Voices for Voice Anonymization. MSc Project, 2026 | UniStuttgart