README.md

---
license: cc-by-4.0
pretty_name: AMI Corpus Mirror (FluidAudio CI)
tags:
  - audio
  - speaker-diarization
  - benchmark
---

# AMI Corpus Mirror

Mirror of the subset of the [AMI Meeting Corpus](https://groups.inf.ed.ac.uk/ami/corpus/) used by
[FluidAudio](https://github.com/FluidInference/FluidAudio) diarization benchmarks. Hosted here so CI and
local benchmark runs do not depend on the availability of the upstream `groups.inf.ed.ac.uk` server
(see [FluidAudio#752](https://github.com/FluidInference/FluidAudio/issues/752)).

## Contents

- `annotations/ami_public_manual_1.6.2.zip` — AMI public manual annotations v1.6.2
  (repackaged from the official archive; identical content, including `segments/`, `words/`,
  `corpusResources/meetings.xml`)
- `sdm/{meeting}.Mix-Headset.wav` — audio for the official 16-meeting AMI-SDM evaluation split
  (EN2002, ES2004, IS1009, TS3003 × a–d), as fetched from the upstream AMI corpus mirror

## License & attribution

The AMI Meeting Corpus is distributed under the
[Creative Commons Attribution 4.0 International (CC BY 4.0)](https://creativecommons.org/licenses/by/4.0/)
license by the AMI Consortium / University of Edinburgh. This repository redistributes it unchanged,
with attribution, under the same license. See `annotations` `LICENCE.txt` inside the zip.

If you use this data, cite:

> Jean Carletta et al., "The AMI Meeting Corpus: A Pre-announcement", MLMI 2005.

## Usage in FluidAudio

`swift run fluidaudiocli diarization-benchmark --auto-download` fetches from this mirror first and
falls back to the upstream Edinburgh server.