OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Audio & Voice AI

Vocal Remover

Vocal Remover is an open-source Python deep-learning application that separates an input song into vocal and instrumental audio tracks. It includes inference commands for CPU and GPU use and a training script for user-provided paired audio datasets.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Vocal Remover — OpenViST project profile in Audio & Voice AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Vocal Remover when you want a local, command-line workflow for separating a single audio file into vocal and instrumental outputs, or when you have paired audio data for the documented training workflow.

Who it is for

  • People who want separate vocal and backing instrumental tracks from a song.
  • Python users who want to run audio-separation inference locally on CPU or GPU.
  • Users with paired instrumental and mixture audio who want to train a model.

Documented uses

  • Extract instrumental and vocal WAV files from an input audio file.
  • Choose between CPU inference and GPU inference.
  • Optionally run inference with test-time augmentation.
  • Organize paired mixtures and instrumentals and invoke the model-training script.
Before you choose

Requirements & limits

The project is published as a public GitHub repository under the MIT license. Its README describes downloading a release and installing PyTorch and the listed Python packages.

  • Inference can be run on CPU or with a selected GPU device.
  • TTA is an optional inference setting.
  • Postprocessing is marked experimental by the project.
  • Training requires paired instrumental and mixture files arranged in separate directories.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up a dataset and start model training

Arrange paired instrumental and mixture files in the documented layout and invoke the training script.

Before you start

  • Install the project and its dependencies as described in the README.
  • Prepare paired audio files for the instruments and mixtures directories.

Steps

  1. Create a dataset directory with instruments and mixtures subdirectories.

  2. Place instrumental tracks in instruments and their corresponding mixtures in mixtures.

  3. Run the command below.

    Documented command
    python train.py --dataset path/to/dataset --mixup_rate 0.5 --reduction_rate 0.5 --gpu 0

Expected outcome

The training script is invoked with the specified dataset and parameters.

Guide limitations

  • The documentation does not explain how to prepare or validate paired data, or specify expected model quality.
  • The documented training command selects GPU device 0.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
tsurumeso/vocal-remover · Official homepage title · retained in its source language
How we review sourcesSuggest a correction