yigagilbert/luganda-speech-cv-yogera

Luganda Speech Data — Common Voice + Yogera (lug)

This repository contains two Luganda speech subsets

|---|---|---| | common-voice-sample-packed-lug/ | common-voice-sample-packed-lug | Mozilla Common Voice | | makerere-yogera-lug/ | makerere-yogera-lug | Makerere AI Lab — Yogera Dataset |

No audio or metadata has been modified — this repo is a filtered re-packaging containing only the Luganda (lug) portions of those two upstream sources, for convenience in downstream ASR/S2ST work. All rights and licensing terms remain with the original data providers as described below.

---

Licensing

1. makerere-yogera-lug/ — Yogera Dataset (Makerere AI Lab)

• License: Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)

• Source of license: AI-Lab-Makerere/Yogera-Dataset-Metadata — LICENSE

• What this means for you as a user of this repo:

  • You may share (copy, redistribute) and adapt (remix, transform, build upon) this data, including for commercial purposes.
  • You must give appropriate credit to Makerere AI Lab / the Yogera Dataset, link to the CC BY-SA 4.0 license, and indicate if changes were made.
  • ShareAlike: if you remix or build upon this data, your derivative must be distributed under the same CC BY-SA 4.0 license.
  • No additional legal or technological restrictions may be placed on others' use of the material beyond what the license permits.
  • The data is provided with no warranty.

  Suggested attribution line:

  Contains data from the Yogera Dataset, Makerere AI Lab, licensed under CC BY-SA 4.0.

2. common-voice-sample-packed-lug/ — Mozilla Common Voice

• Repository license file referenced: common-voice/common-voice — LICENSE (Mozilla Public License 2.0)

• Important distinction: the LICENSE file at that URL covers the Common Voice platform/codebase (the web app and tooling Mozilla built to collect voice data), not the recorded voice clips themselves. Mozilla separately dedicates the actual Common Voice audio + transcript dataset to the public domain under CC0 1.0 Universal, as stated on the Common Voice dataset page and its Terms of Use.

• Practical implication:

  • The audio clips and transcripts in common-voice-sample-packed-lug/ are effectively public domain (CC0) — no attribution is legally required, though crediting Mozilla Common Voice is good practice.
  • MPL 2.0 would only be relevant if you were redistributing Common Voice's software/code, which is not part of this repo.
  • ⚠️ Recommended: before publishing, verify with Sunbird AI how this subset was packaged (e.g. whether any Common Voice metadata schema/code was bundled in), since that could reintroduce an MPL 2.0 component. Absent that, treat the clips as CC0.

  Suggested attribution line:

  Contains data from the Mozilla Common Voice dataset (CC0 1.0), https://commonvoice.mozilla.org

---

Summary table

Subset	License (data)	Commercial use	Attribution required	ShareAlike required
Yogera (lug)	CC BY-SA 4.0	✅ Yes	✅ Yes	✅ Yes (derivatives must stay CC BY-SA 4.0)
Common Voice (lug)	CC0 1.0 (public domain)	✅ Yes	Not required (recommended)	❌ No

Net effect on this repo: because it combines a CC BY-SA 4.0 subset with a CC0 subset, and CC0 imposes no restrictions, this repository as a whole should be treated as CC BY-SA 4.0 for any use that touches the Yogera folder, unless you keep the two folders and their licenses clearly separated when redistributing.

---

Disclaimer

No warranties are made about the data. Licensing information above reflects what is stated at the linked upstream sources as of the date above — always check the original LICENSE pages for the current, authoritative terms before redistributing.