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

Hs Tasnet

Hs Tasnet is a Python and PyTorch implementation of Hybrid Spectrogram-TasNet for separating mixed stereo audio into multiple tracks. The library includes model inference, a training utility, checkpoint-loading methods and streaming inference examples. It is intended for developers integrating source separation into audio software or experimenting with model training, not a complete end-user editing application.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Hs Tasnet — 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 HS-TasNet when you need a Python model implementation to integrate or experiment with stereo music source separation, and can build the surrounding application or workflow yourself.

Who it is for

  • Audio-software developers adding track separation to an application.
  • Researchers or engineers experimenting with music-separation models.
  • Teams training a customized model with an in-house dataset.

Documented uses

  • Pass stereo audio to a model and receive separated tracks.
  • Train the model using MusDB or an in-house dataset.
  • Load a saved checkpoint for inference.
  • Prototype streaming audio separation and request selected source indices.
Before you choose

Requirements & limits

The repository is public and identifies an MIT license. Its README documents installation with pip.

  • This is a library rather than a finished audio-editing application; integrating it into a user-facing tool requires additional development.
  • The usage example returns four separated tracks, but does not label them by instrument.
  • Training setup uses a separate dependency-install script and uv; distributed training adds an Accelerate configuration step.
  • The README documents replacing the default LSTM or GRU with another recurrent module that follows the specified interface.

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

Prepare and start the documented training script

Set up the listed dependencies and launch the repository’s training script.

Before you start

  • The repository source.
  • uv installed with pip.

Steps

  1. Run the command below to install dependencies.

    Documented command
    sh scripts/install.sh

  2. Install uv with the command below if it is not already available.

    Documented command
    pip install uv

  3. Run the command below to train a newly initialized model on a small subset of MusDB.

    Documented command
    uv run train.py

Expected outcome

The documented training script is launched on a small MusDB subset.

Guide limitations

  • The README does not provide a guaranteed training result; its instruction is to check that the loss goes down.
  • Distributed training requires running accelerate config first.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Hs TasnetPass stereo audio to a model and receive separated tracks.CLIThe repository is public and identifies an MIT license. Its README documents installation with pip.
BluettsGenerate speech audio from text in a supported language.CLI, Linux, macOS, WindowsThe repository is public and identifies the project license as MIT.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
12 October 2026
Original website title
lucidrains/HS-TasNet · Official homepage title · retained in its source language
How we review sourcesSuggest a correction