Research and Production Oriented Speaker Verification, Recognition and Diarization Toolkit

This project uses a variety of advanced voiceprint recognition models such as EcapaTdnn, ResNetSE, ERes2Net, CAM++, etc. It is not excluded that more models will be supported in the future. At the same time, this project also supports MelSpectrogram, Spectrogram data preprocessing methods

Unofficial reimplementation of ECAPA-TDNN for speaker recognition (EER=0.86 for Vox1_O when train only in Vox2)

The Pytorch implementation of sound classification supports EcapaTdnn, PANNS, TDNN, Res2Net, ResNetSE and other models, as well as a variety of preprocessing methods.

本项目使用了EcapaTdnn、ResNetSE、ERes2Net、CAM++等多种先进的声纹识别模型，同时本项目也支持了MelSpectrogram、Spectrogram、MFCC、Fbank等多种数据预处理方法

基于PaddlePaddle实现的音频分类，支持EcapaTdnn、PANNS、TDNN、Res2Net、ResNetSE等各种模型，还有多种预处理方法

Verifying the identity of a person from characteristics of the voice independent from language via NVIDIA NeMo models (ECAPA-TDNN, SpeakerNet, TitaNet-L).

针对CN-Celeb数据集的基于ECAPA-TDNN的说话人识别的pytorch实现

Speaker verification task with ECAPA-TDNN model (trained on Persian dataset)

A multimodal SER project combining BERT and ECAPA-TDNN with cross-attention-based fusion on the IEMOCAP dataset.

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

Speaker verification of virtual assistants using ECAPA-TDNN model from SpeechBrain toolkit and transfer learning approach emphasizing on inter and intra comparision (text independent and dependent).

Multi-modal biometric authentication system combining face recognition (FaceNet), speaker verification (ECAPA-TDNN), and speech-to-text with liveness detection. Built with PyTorch. 99.5% test accuracy.

Speaker anonymization & verification with ECAPA and Integrated Gradients

A text-independent speaker verification system built with ECAPA-TDNN (Emphasized Channel Attention, Propagation and Aggregation in Time Delay Neural Network) from SpeechBrain.

This project is a Voice Identification System built using Python, leveraging SpeechBrain and ECAPA-TDNN for speaker verification. The system identifies users by comparing their voice embeddings with stored data, providing a secure and efficient method for user recognition.

ECAPA-TDNN + Integrated Gradients to explain speaker verification and the impact of pitch-shift anonymization on LibriSpeech (with EER and IG heatmaps)

CryCeleb2023 experiments

ECAPA-TDNN leakage analysis across +1…+4 semitones: feature & waveform IG heatmaps, LibriSpeech evaluation, EER, failures/near-misses, and frame-aligned time-frequency leakage metrics.