Skip to content

Dataset · angeluriot

French instructions and dialogues

angeluriot/french_instruct

276k instruction-response pairs and multi-turn dialogues in French for language model training.

What's inside

Contains 276 thousand user-assistant dialogues (including multi-turn conversations) totaling about 85 million tokens entirely in French. Each document includes metadata indicating whether the text was human-written or model-generated. The dataset is also mirrored on GitHub.

What you can do with it

  • SFT fine-tuning and adaptation of LLMs for French language tasks
  • Training classifiers to detect AI-generated text
  • Developing conversational assistants and question-answering search systems
Tasks
Question answering, Text generation, Text classification, Token classification
Languages
French
Size
100K–1M rows
Downloads this month
868
Likes
21

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("angeluriot/french_instruct")