# karolpiczak/ESC-50

ESC-50: Dataset for Environmental Sound Classification

Repository: https://github.com/karolpiczak/ESC-50
Canonical: https://ross.abutalabs.com/products/esc-50
Language: Python
License: NOASSERTION
License Family: other
Topics: dataset, environmental-sounds, audio
Last push: 2024-03-20T15:07:26+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 4158, "days_push": 896, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1859, forks 326 (observed 2026-08-28T04:05:45.091058+00:00)

## What it is
ESC-50 is a labeled dataset of 2000 five-second environmental audio recordings organized into 50 classes for benchmarking environmental sound classification. It includes prearranged cross-validation folds and clips sourced from Freesound.org field recordings.

## Use cases
- benchmark environmental sound classification models
- train an audio classifier to recognize sounds like dogs barking or rain
- evaluate machine learning methods on environmental audio
- get a labeled dataset for audio classification research
- compare audio classification results with published baselines
- set up cross-validation folds for sound recognition experiments

## When to choose
- you need a standard benchmark for environmental sound classification
- you want a small, manageable labeled audio dataset for research or teaching
- you need prearranged cross-validation folds for comparable results

## When to avoid
- you need large-scale audio data beyond 2000 clips
- you require a permissive commercial-use license (it is CC BY-NC)
- you need speech or music classification data rather than environmental sounds

## Facets
- artifact type: dataset
- maturity: stable
- function: audio-processing, machine-learning, benchmarking
- domain: machine-learning, data-science
- platform: cross-platform, python
- tags: environmental-sound-classification, audio-dataset, benchmark, freesound, audio-classification, audio

## Member repositories
- karolpiczak/ESC-50 (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:45.091058+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 2026-08-30T03:16:00.907066+00:00, confidence not recorded.
  - readme: https://github.com/karolpiczak/ESC-50 (fetched 2026-08-28T04:05:45.091058+00:00, sha c9fd646d14e5)
- Data as of 2026-08-30T08:39:29.467469+00:00.