metadata

license: mit
task_categories:
  - text-generation
language:
  - br
size_categories:
  - 10K<n<100K

Goulenn

A Breton Instructions Dataset called Goulenn (meaning "Question" in Breton).
Direct translation of the jpacifico/French-Alpaca-dataset-Instruct-110K by Jonathan Pacifico.
For now only 50k samples have been translated, 110k version soo to come...
Generation details available on the GweLLM Github repository.

Sample test code:

from datasets import load_dataset

dataset = load_dataset( path="amurienne/Goulenn-Alpaca-Instruct-50k",
                        split="train")

print(f"dataset infos:\n{dataset}")
print(f"dataset first sample:\n{dataset[0]}")

Created by Albert Murienne, 2025
Please credit my name if you use this dataset in your project.