OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Bs Roformer

Bs Roformer is an MIT-licensed Python and PyTorch implementation of Band Split RoFormer, an attention-based music source-separation model that applies axial attention across frequency bands and time. It includes BSRoformer, MelBandRoformer, and a FlowBSRoformer variant, plus documented support for stereo training and multiple-stem output.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Bs Roformer — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a project needs an implementation-level foundation for music source separation and the team can provide the surrounding training, inference, or application workflow.

Who it is for

  • Python and PyTorch developers building self-hosted audio tools.
  • Researchers studying attention-based music source separation.
  • Audio developers developing stereo or multi-stem processing workflows.

Documented uses

  • Separate music into multiple audio stems.
  • Train or study a Band Split RoFormer implementation.
  • Experiment with rotary positional encoding and its successor option.
  • Develop a stereo-aware source-separation component.
  • Compare standard, Mel-Band, and flow-matching model variants.
Before you choose

Requirements & limits

The repository is publicly accessible and identified with an MIT license. The documented installation uses pip; no hosted service, pricing, account requirement, or packaged application is documented.

  • The implementation uses Python and PyTorch.
  • Stereo training and multiple-stem output are explicitly supported.
  • The repository includes separate BS-Roformer, Mel-Band Roformer, and FlowBSRoformer model paths.
  • The examples cover tensor-based training and output generation rather than a complete end-user application.
  • A batch stem-extraction workflow is Likely compatible, but the complete batch application is not documented.

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 the BS-RoFormer Python implementation

Install the package and instantiate a documented separation model for training-oriented experimentation.

Before you start

  • Python environment with PyTorch available.
  • Access to the public MIT-licensed repository.
  • Audio tensors shaped as shown in the repository examples.

Steps

  1. Install the package with the documented pip command.

  2. Import BSRoformer from the bs_roformer package.

  3. Instantiate the model with documented dimensions and transformer depths.

  4. Pass input and target tensors to calculate loss, then call backward during training.

  5. After training, pass input audio to the model to obtain output.

Expected outcome

A configured BS-RoFormer model suitable for the documented training and output workflow.

Guide limitations

  • The example does not provide pretrained weights or a complete production inference application.
  • The repository does not document a full batch-processing interface or user interface.
Documentation 1 ↗
Evidence and transparency

Sources & review

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