{"id":17082043,"url":"https://github.com/fakufaku/fast_bss_eval","last_synced_at":"2025-04-09T12:07:38.643Z","repository":{"id":37686309,"uuid":"416243695","full_name":"fakufaku/fast_bss_eval","owner":"fakufaku","description":"A fast implementation of bss_eval metrics for blind source separation","archived":false,"fork":false,"pushed_at":"2022-06-07T01:37:15.000Z","size":1162,"stargazers_count":126,"open_issues_count":4,"forks_count":8,"subscribers_count":4,"default_branch":"main","last_synced_at":"2024-06-11T19:59:07.987Z","etag":null,"topics":["audio-processing","blind-source-separation","bss","evaluation-metrics","fast"],"latest_commit_sha":null,"homepage":"https://fast-bss-eval.readthedocs.io/en/latest/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/fakufaku.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.rst","contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2021-10-12T08:11:25.000Z","updated_at":"2024-04-13T08:48:36.000Z","dependencies_parsed_at":"2022-07-09T03:15:06.406Z","dependency_job_id":null,"html_url":"https://github.com/fakufaku/fast_bss_eval","commit_stats":null,"previous_names":[],"tags_count":7,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/fakufaku%2Ffast_bss_eval","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/fakufaku%2Ffast_bss_eval/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/fakufaku%2Ffast_bss_eval/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/fakufaku%2Ffast_bss_eval/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/fakufaku","download_url":"https://codeload.github.com/fakufaku/fast_bss_eval/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248036067,"owners_count":21037092,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["audio-processing","blind-source-separation","bss","evaluation-metrics","fast"],"created_at":"2024-10-14T12:59:08.318Z","updated_at":"2025-04-09T12:07:38.626Z","avatar_url":"https://github.com/fakufaku.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"fast\\_bss\\_eval\n==============\n\n[![PyPI version](https://badge.fury.io/py/fast-bss-eval.svg)](https://badge.fury.io/py/fast-bss-eval)\n[![Documentation Status](https://readthedocs.org/projects/fast-bss-eval/badge/?version=latest)](https://fast-bss-eval.readthedocs.io/en/latest/?badge=latest)\n[![black](https://github.com/fakufaku/fast_bss_eval/actions/workflows/lint.yml/badge.svg?branch=main)](https://github.com/fakufaku/fast_bss_eval/actions/workflows/lint.yml)\n[![tests](https://github.com/fakufaku/fast_bss_eval/actions/workflows/pythonpackage.yml/badge.svg)](https://github.com/fakufaku/fast_bss_eval/actions/workflows/pythonpackage.yml)\n[![codecov](https://codecov.io/gh/fakufaku/fast_bss_eval/branch/main/graph/badge.svg?token=N0OF1R2OWG)](https://codecov.io/gh/fakufaku/fast_bss_eval)\n\n\u003e Do you have a zillion BSS audio files to process and it is taking days ?\n\u003e Is your simulation never ending ?\n\u003e\n\u003e Fear no more! `fast_bss_eval` is here to help **you!**\n\n`fast_bss_eval` is a fast implementation of the bss\\_eval metrics for the\nevaluation of blind source separation.  Our implementation of the bss\\_eval\nmetrics has the following advantages compared to other existing ones.\n\n* seamlessly works with **both** [numpy](https://numpy.org/) arrays and [pytorch](https://pytorch.org) tensors\n* very fast\n* can be even faster by using an iterative solver (add `use_cg_iter=10` option to the function call)\n* differentiable via pytorch\n* can run on GPU via pytorch\n\nAuthor\n------\n\n* [Robin Scheibler](robin.scheibler@linecorp.com)\n\nQuick Start\n-----------\n\n### Install\n\n```bash\n# from pypi\npip install fast-bss-eval\n\n# or from source\ngit clone https://github.com/fakufaku/fast_bss_eval\ncd fast_bss_eval\npip install -e .\n```\n\n### Use\n\nAssuming you have multichannel signals for the estmated and reference sources\nstored in wav format files names ``my_estimate_file.wav`` and\n``my_reference_file.wav``, respectively, you can quickly evaluate the bss_eval\nmetrics as follows.\n\n\n```python\nfrom scipy.io import wavfile\nimport fast_bss_eval\n\n# open the files, we assume the sampling rate is known\n# to be the same\nfs, ref = wavfile.read(\"my_reference_file.wav\")\n_, est = wavfile.read(\"my_estimate_file.wav\")\n\n# compute the metrics\nsdr, sir, sar, perm = fast_bss_eval.bss_eval_sources(ref.T, est.T)\n```\n\nBenchmark\n---------\n\nThis package is significantly faster than other packages that also allow\nto compute bss\\_eval metrics such as [mir\\_eval](https://github.com/craffel/mir_eval) or [sigsep/bsseval](https://github.com/sigsep/bsseval).\nWe did a benchmark using numpy/torch, single/double precision floating point\narithmetic (fp32/fp64), and using either Gaussian elimination or a [conjugate\ngradient descent](https://en.wikipedia.org/wiki/Conjugate_gradient_method)\n(solve/CGD10).\n\n\u003cimg src=\"./docs/figures/channels_vs_runtime.png\"\u003e\n\n\nCitation\n--------\n\nIf you use this package in your own research, please cite [our paper](https://arxiv.org/abs/2110.06440) describing it.\n\n```bibtex\n@misc{scheibler_sdr_2021,\n  title={SDR --- Medium Rare with Fast Computations},\n  author={Robin Scheibler},\n  year={2021},\n  eprint={2110.06440},\n  archivePrefix={arXiv},\n  primaryClass={eess.AS}\n}\n```\n\n\nLicense\n-------\n\n2021 (c) Robin Scheibler, LINE Corporation\n\nAll of this code is released under [MIT License](https://opensource.org/licenses/MIT) with the exception of `fast_bss_eval/torch/hungarian.py` which is under [3-clause BSD License](https://opensource.org/licenses/BSD-3-Clause).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ffakufaku%2Ffast_bss_eval","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Ffakufaku%2Ffast_bss_eval","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Ffakufaku%2Ffast_bss_eval/lists"}