News
Whisper large v3 Tatar–02/10/2026
openai/whisper-large-v3, fully fine-tuned on Common Voice Scripted Speech 27.0 Tatar (CC0). It writes Tatar in Cyrillic, with capital letters and punctuation, so its output can be used as subtitles with little editing.
imlaTAT 0.7–29/09/2026
OsmanLat (Ottoman to Latin Turkish model) retrained on 3+k lines from seven printed editions of 1906 to 1935. Reads Arabic-script Old Tatar and outputs Cyrillic. On held-out lines the character error is 3.4 % and the word error is 17 %. With the lexicon decoder they are 2.9 % and 12.6 %. Runs offline, 0.1 to 0.2 s per line.
TATlit–31/08/2026
TATlit is a 478M-parameter, from-scratch base language model for literary Tatar. Pretrained on Tatar and Qypchaq siblings, then solely on Tatar. On FLORES, BOUQuET and TatBLiMP it scores best among the tested models, from Tatar-specialized ones to frontier models of up to 120B parameters.
Tatar (Cyrillic) OCR model for Tesseract 5–26/06/2026
A Tesseract 5 LSTM model that reads printed Tatar in Cyrillic, including the six Tatar-specific letters ә ө ү җ ң һ. Trained for digitizing Tatar books and periodicals.
TatBLiMP–19/06/2026
TatBLiMP measures Tatar (tt, Cyrillic) morphosyntax with minimal pairs. It follows BLiMP (Warstadt et al. 2020) and adapts most of TurBLiMP's phenomena to Tatar.
Whisper Fine-tuned Model–02/05/2026
This model is fine-tuned from openai/whisper-small.
Whisper Fine-tuned ASR–30/04/2026
This is a fine-tuned version of OpenAI Whisper for speech recognition.
F5-TTS Fine-tuned Model–30/04/2026
This model is fine-tuned from F5TTS_v1_Base using TatarTTS dataset.
Tatar OCR Benchmark–31/03/2026
This dataset is a page-level OCR benchmark for 251 Tatar documents. Full original page images are included.
Tatar Youtube Audios–10/08/2025
Studio Dub Recordings–29/06/2025
Studio Dub Recordings is a speech dataset consisting of high-quality studio recordings of dubbed audio in Tatar language.
Tatar Audiobooks–04/05/2025
170 hours of aligned audiobooks taken from tatkniga.ru. There are 4 speakers with 17+ hours of audio and 20 speakers in total.
Plain text of Wikipedia–02/05/2025
This dataset is a plain text version of pages from wikipedia.org spaces for several languages (Tatar, Bashkir, Sakha). The text is without HTML tags nor wiki templates.
Volga Fairytales TTS–30/04/2025
Volga Fairytales TTS is a speech dataset containing audio recordings of traditional fairytales in several languages of the Volga region.
TatAlpaca–22/02/2025
Dataset of Gemini-generated instructions in Tatar language.
Tatar-exams–13/11/2024
Tatar-exams - part of TUMLU: A Unified and Native Language Understanding Benchmark for Turkic Languages. This repo contains the Tatar language part of TUMLU-mini dataset from TUMLU paper.
Tatar-English Parallel Corpus–19/09/2024
The Tatar-English Parallel Corpus is a collection of parallel sentences in Tatar and English. It is designed to facilitate research and development in machine translation, natural language processing (NLP), and multilingual studies.
Tatar Language Dictionaries–16/05/2024
This dataset, hosted by Yasalma, is a curated collection of multiple Tatar language dictionaries in Parquet format. Each file represents a single dictionary, making this dataset a comprehensive resource for linguistic research and Tatar language processing.
Tatar Books Collection (Cyrillic)–29/04/2024
This dataset, hosted by Yasalma, is a curated collection of 497 Tatar books in Parquet format. The texts are in Cyrillic script, making this dataset ideal for linguistic research, language modeling, and other NLP applications in the Tatar language.
TatarCrawl–25/04/2024
In an effort to democratize research on low-resource languages, we release TatarCrawl dataset, a web news corpus consisting of materials from nearly 15 unique sources in the Tatar Language.
Tatar-Russian parallel corpus–25/04/2024
The part of the dataset with the label "source" = "ipsan" has been provided by the Institute of Applied Semiotics of Tatarstan Academy of Sciences. Around 600k parallel sentences.