Metraipolis › archive › dc_09756ef436d4
ufakzeka-kulak: Building and Evaluating a 31.6M-Parameter Turkish Speech Recognizer from Scratch — arXiv:2611.03605
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.03605
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T11:11:35.747Z
- SHA-256 of the archived text
5c99d2f396a50536691a19c6b4983b4304b0fd3f14313f0a69a97ab27f676697- Stable URL of this page
https://metraipolis.com/documento/dc_09756ef436d4
How to cite
ufakzeka-kulak: Building and Evaluating a 31.6M-Parameter Turkish Speech Recognizer from Scratch — arXiv:2611.03605 — arXiv (en), https://arxiv.org/abs/2611.03605 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_09756ef436d4 (Metraipolis archive dc_09756ef436d4, sha256 5c99d2f396a50536691a19c6b4983b4304b0fd3f14313f0a69a97ab27f676697)
BibTeX
@misc{dc09756ef436d4,
title = {ufakzeka-kulak: Building and Evaluating a 31.6M-Parameter Turkish Speech Recognizer from Scratch — arXiv:2611.03605},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.03605},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_09756ef436d4; sha256 5c99d2f396a50536691a19c6b4983b4304b0fd3f14313f0a69a97ab27f676697},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.03605. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
ufakzeka-kulak is an open Turkish speech-to-text model of 31.6M parameters, a FastConformer encoder with a token-and-duration transducer (TDT) and a CTC head, trained from scratch and released under Apache-2.0. It writes cased, punctuated Turkish with numbers as words. Its 4,977 training hours come from sources whose licences allow commercial use; 281 hours are human-transcribed, and for the other 4,696 the training text is an uploader subtitle or a machine transcript, kept only where a second, independent reading of the audio agrees with it. The recipe, checkpoint, labelling and export choices were made on development data, and the test sets were scored once, in a release gate fixed before training. With greedy TDT decoding the normalised word error rate (WER) is 0.0921 (95 percent interval 0.0792 to 0.1055) on FLEURS Turkish test and 0.0916 (0.0873 to 0.0958) on Common Voice 27.0 Turkish test, lower than Whisper base and Whisper small on both sets with paired intervals wholly below zero; Whisper large-v3 (greedy) scores 0.0604 and 0.0767. On conversational speech the WER is 0.3807. We also report punctuation and casing, degraded audio, and how closely each shipped export reproduces the trained model. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Audio and Speech Processing (eess.AS). Pagina del lavoro: https://arxiv.org/abs/2611.03605 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-03T10:59:52.369Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).