ufakzeka-kulak：从零构建并评估 31.6M 参数土耳其语语音识别模型

View PDF HTML (experimental)

Abstract: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.

Comments:	9 pages (text on pages 1 to 8, references on pages 8 and 9). Model, card, attribution, code and demo: this https URL, this https URL, this https URL
Subjects:	Computation and Language (cs.CL); Audio and Speech Processing (eess.AS)
ACM classes:	I.2.7
Cite as:	arXiv:2611.03605 [cs.CL]
	(or arXiv:2611.03605v1 [cs.CL] for this version)
	https://doi.org/10.48550/arXiv.2611.03605 arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Sait Furkan Teke [view email]
[v1] Mon, 2 Nov 2026 18:42:32 UTC (15 KB)