This project implements a voiceprint/speaker recognition system using various deep learning models including EcapaTdnn, ResNetSE, ERes2Net, and CAM++. It supports multiple audio preprocessing methods such as MelSpectrogram, Spectrogram, MFCC, and Fbank. The system uses ArcFace loss (AAMLoss) along with other loss functions like AM loss, ARM loss, and CE loss for training speaker embedding networks.