Back to catalog

Qırımtatar TTS

An open text-to-speech model for Crimean Tatar with an interactive demo.

github.com

About this project

  • Converts Crimean Tatar text into speech
  • The interactive version runs on Hugging Face Spaces
  • Source code is published under the MIT license

Qırımtatar TTS Datasets

Language project#Language#Technology#Text to speech#Dataset#Open sourceCrimean Tatars
  • Apache 2.0 license, maintained by developer egorsmkv
  • About 8.8 hours of audio in OPUS 48 kHz format
  • Mirrored on Hugging Face with an interactive demo
Learn more

Crimean Tatar TTS — Sevil

Language project#Language#Science#Architecture#Text to speech#Open sourceCrimean Tatars
  • Trained on 1,566 audio recordings via transfer learning from the multilingual SpeechT5 model
  • Supports both Latin and Cyrillic Crimean Tatar script
  • Validated by native speakers
Learn more

Qirimtatar Embedding Crawlers

Language project#Language#Technology#Open sourceCrimean Tatars
  • Collects texts from open web sources and digital archives
  • Uses Tesseract for text recognition in PDFs and images
  • Publishes the extracted documents separately from the source code
  • Processes the Qirim Junior archives and the texts of Shamil Aladin
Learn more
Муэллиф