Skip to content

Dataset · KeremTT

Balanced English and Turkish chat dataset

KeremTT/FABLE-5.2-MAX-DATABASE

240,000 dialogue examples across 12 domains in English and Turkish for LLM training.

What's inside

The dataset contains 240,000 text lines in English and Turkish, evenly distributed across 12 categories (exactly 20,000 lines each). Data is collected from open-source repositories with permissive licenses and formatted into a unified chat schema. Delivered as 48 JSONL files with 5,000 lines each.

What you can do with it

  • SFT fine-tuning of multilingual LLMs with Turkish and English support
  • Training text generation and question-answering models across domains
  • Building conversational assistants based on open weights
Tasks
Text generation, Question answering
Languages
English, Turkish
Size
100K–1M rows
Downloads this month
700
Likes
9

How to load it

from datasets import load_dataset

ds = load_dataset("KeremTT/FABLE-5.2-MAX-DATABASE")