audio-separator has 482K monthly downloads and 1K GitHub stars. Version 0.44.2 supports Python 3.10+, first published in 2023.
It is stably maintained under a permissive license, last released in May 2026.
It is built on top of audioop-lts, beartype and diffq. It is used by 22 tracked packages, including karaoke-gen, ultimate-rvc and karaoke-prep.
Package Insights
((week_daily_avg - month_daily_avg) / month_daily_avg) * 100Weekly Downloads
GitHub Stars
Downloads by OS
Python Versions
Top Countries
Dependencies
- audioop-lts >=0.2.1
- beartype <0.19.0,>=0.18.5
- diffq >=0.2
- diffq-fixed >=0.2
- einops >=0.7
- julius >=0.2
- librosa >=0.10
- ml-collections
- numpy >=2
- onnx-weekly
- onnx2torch-py313 >=1.6
- pydub >=0.25
- pyyaml
- requests >=2
- resampy >=0.4
- rotary-embedding-torch <0.7.0,>=0.6.1
- samplerate ==0.1.0
- scipy <2.0.0,>=1.13.0
- six >=1.16
- soundfile >=0.12
- torch >=2.3
- tqdm
4 optional dependencies
- onnxruntime[cpu]
- onnxruntime-directml[dml]
- onnxruntime-gpu[gpu]
- torch-directml[dml]