# bytedance/music_source_separation

Repository: https://github.com/bytedance/music_source_separation
Canonical: https://ross.abutalabs.com/products/music_source_separation
Language: Python
License: NOASSERTION
License Family: other
Topics: research
Archived: true
Last push: 2024-04-18T06:24:38+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1933, "days_push": 867, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1383, forks 200 (observed 2026-08-28T04:04:34.623486+00:00)

## What it is
A PyTorch implementation of music source separation that splits songs into sources like vocals and accompaniment. It ships pretrained checkpoints and a CLI for inference, plus full training scripts to build custom separation systems.

## Use cases
- separate vocals from accompaniment in a song
- remove vocals from mp3 tracks
- batch separate a directory of audio files
- train a custom source separation model on MUSDB18
- train speech enhancement or instrument separation systems
- integrate a separation model into my own Python program

## When to choose
- you need vocal/accompaniment separation with pretrained checkpoints
- you want a reproducible PyTorch training pipeline for separation tasks
- you need batch CLI processing of audio files

## When to avoid
- you need real-time or streaming separation
- you need many source types beyond vocals/accompaniment out of the box
- you need a maintained GUI or production-grade service

## Facets
- artifact type: library
- maturity: maintenance
- function: audio-processing, machine-learning, deep-learning, cli
- domain: machine-learning, media
- platform: python, cross-platform
- tags: music-source-separation, pytorch, vocal-removal, audio-separation, musdb18, speech-enhancement, audio, linux, macos

## Member repositories
- bytedance/music_source_separation (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:34.623486+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:40:01.994542+00:00, confidence not recorded.
  - readme: https://github.com/bytedance/music_source_separation (fetched 2026-08-28T04:04:34.623486+00:00, sha a92e5fb92c73)
- Data as of 2026-08-30T08:39:29.467469+00:00.