metadata
license: mit
language:
- tr
- en
Details
This is a mixed ~600M tokens high quality dataset consisting of personally scraped & publicly available data with <600 tokens strictly in Turkish & English.
Please credit if you use it in your training.