AIbase
EN
Home
Convtasnet Libri2Mix Sepclean 16k
C

Convtasnet Libri2Mix Sepclean 16k

Developed by JorisCos
This is a ConvTasNet model trained based on the Asteroid framework, specifically designed for audio separation tasks, trained on the sep_clean task of the Libri2Mix dataset.
Sound Separation
PyTorch
#Speech Separation#High-Fidelity Audio#ConvTasNet Architecture
Downloads 13.38k
Release Time : 3/2/2022

Model Overview

This model is primarily used to separate clean speech signals from mixed audio, suitable for speech enhancement and audio processing scenarios.

Model Features

Efficient Audio Separation
Utilizes the ConvTasNet architecture to achieve high-quality speech separation
Optimized Training Configuration
Employs carefully designed filter banks and mask network configurations to enhance separation performance
Excellent Performance Metrics
Demonstrates outstanding performance on the Libri2Mix test set, achieving an SI-SDR of 15.24dB

Model Capabilities

Speech Separation
Audio Enhancement
Multi-Speaker Separation

Use Cases

Speech Processing
Meeting Recording Separation
Separates individual speaker voices from multi-person meeting recordings
SI-SDR improvement of 15.24dB
Speech Enhancement
Extracts clear speech from background noise
STOI improvement of 0.226
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase