JudSacr/

CFP

This dataset contains native French prompt data (in the sense that it is not a translation of an English dataset) and manually clean.

from datasets import load_dataset dataset = load_dataset("CATIE-AQ/CFP")

	All data (56,277 questions)




	
	Tasks covered:

faq: 16,668 (29.62%) (if faq is considered open_qa, then 56.80% of data is open_qa)

open_qa: 15,298 (27.18%)

mrc: 900 (1.60%)

qam: 5,000 (8.88%)… See the full description on the dataset page: https://huggingface.co/datasets/JudSacr/CFP.