---
id: mir-eval
version: "0.8.2"
license: MIT
license_treatment: permissive
maintenance: active
---

# mir-eval — Common metrics for common audio/music processing tasks.

License: permissive · Maintenance: active · Downloads: 229.7K/mo

## What it is and what it does

mir_eval is a Python library that implements standard evaluation metrics used in music and audio information retrieval research. It provides reference implementations of common accuracy measures for audio signal processing problems. The library depends on numpy, scipy, and decorator, making it straightforward to integrate into research workflows and evaluation pipelines.

The package is designed for researchers and practitioners who need to benchmark audio processing algorithms against established metrics. Rather than reimplementing evaluation logic, users can call mir_eval's functions to compute metrics consistently and reproducibly, which is especially valuable when comparing results across papers or validating new methods against published baselines.

Use it for:

- Evaluate audio processing algorithms by computing standard metrics against ground truth annotations.
- Benchmark music processing systems using standardized accuracy measures from the mir_eval library.
- Validate audio analysis results by computing evaluation metrics across research projects.
- Compare results across papers using transparent, reference implementations of common MIR metrics.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

mir_eval computes standard accuracy metrics for music and audio information retrieval tasks, enabling transparent evaluation of audio processing algorithms.

Yes. mir_eval is actively maintained, has no known vulnerabilities, installs with low friction, and provides standard reference implementations for music information retrieval metrics. Install it if you are evaluating audio or music processing algorithms and need reproducible metric computation.

## Install

pip install mir-eval
uv add mir-eval
poetry add mir-eval

## Installing mir-eval

Before you install: Low friction installation with three stable dependencies (numpy, scipy, decorator). Active maintenance with recent commits and a 710-star repository; last release was 2025-02-25.

License in practice: MIT license is permissive, allowing use in commercial and proprietary projects with minimal restrictions beyond attribution.

Quickstart:

pip install mir-eval

import mir_eval

# Call mir_eval functions to compute metrics for your audio processing task

Verify before relying:

- Whether the library supports specific MIR tasks like beat tracking, chord recognition, and source separation evaluation.
- Whether mir_eval provides F-measure, precision, recall, or other specific metric types.

## Package facts

- License: MIT (permissive)
- Python support: unspecified
- Install friction: low
- Maintenance: active
- Downloads: 229.7K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

music information retrieval metrics, audio evaluation scores, mir evaluation library, audio signal processing evaluation, music processing benchmarking, music-information-retrieval, audio-evaluation, research-metrics

[View on SkillFed](https://skillfed.io/packages/mir-eval) · [View on PyPI](https://pypi.org/project/mir-eval/)