{
 "id": "davidilag/FPSC",
 "owner": "davidilag",
 "name": "FPSC",
 "downloads": 1296,
 "likes": 0,
 "last_modified": "2026-05-13T15:48:01.000Z",
 "stated_licence": "cc-by-4.0",
 "stated_from": "tag",
 "licence_name": null,
 "licence_link": null,
 "file_licence": null,
 "licence_files": [],
 "sources": [],
 "gated": false,
 "tasks": [
  "automatic-speech-recognition"
 ],
 "citation": "@inproceedings{iLag2026FPSC,\r\n  author = {Dávid í Lág and Barbara Scalvini and Carlos Mena and Jón Guðnason},\r\n  title = {{FPSC}: A Sustainable Pipeline for Building a Faroese Parliamentary Speech Corpus},\r\n  booktitle = {Proceedings of the Language Resources and Evaluation Conference (LREC 2026)},\r\n  year = {2026},\r\n  address = {Palma de Mallorca, Spain},\r\n  publisher = {European Language Resources Association (ELRA)},\r\n  institution = {University of the Faroe Islands},\r\n  keywords = {Faroese, Parliamentary Speech, Automatic Speech Recognition, Weakly-Supervised Transcription, Whisper, Wav2Vec2},\r\n  abstract = {We present FPSC, a 1,600-hour Faroese parliamentary speech corpus comprising approximately 89,000 speeches enriched with detailed speaker and linguistic metadata. The corpus was constructed using a sustainable ASR-assisted pipeline combining speech segmentation, multiple Faroese-adapted ASR systems, and ROVER-based consensus voting for weakly supervised transcription. FPSC represents the first large-scale corpus of natural spoken Faroese and provides an open resource for future research in automatic speech recognition and low-resource language technology.}\r\n}",
 "signal": "permissive",
 "reason": "The declared licence (cc-by-4.0) is classified as permissive.",
 "revision": "ce9940cd8a707cfca93fa07c7ac4b35ec8cfc880",
 "stated_evidence": "https://huggingface.co/datasets/davidilag/FPSC/blob/ce9940cd8a707cfca93fa07c7ac4b35ec8cfc880/README.md",
 "file_evidence": "https://huggingface.co/datasets/davidilag/FPSC/tree/ce9940cd8a707cfca93fa07c7ac4b35ec8cfc880",
 "card_nc": false
}