metadata

dataset_info:
  features:
    - name: text
      dtype: string
    - name: token_count
      dtype: int64
  splits:
    - name: train
      num_bytes: 45492446
      num_examples: 57010
  download_size: 26622228
  dataset_size: 45492446
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-sa-3.0
language:
  - ga

Irish Wikipedia - Vicipéid (Gaeilge)

This dataset is generated from the 20240501 wikimedia dump.

The "text" column is obtained by combining the article_title and article_text from each wikipedia article, as follows:

text = f'{article_title}. {article_text}'

Token_count is obtained by using the Llama 3 tokenizer.

License

This dataset is derived from content from Wikipedia, licensed under CC BY-SA 3.0. The original content is available at https://ga.wikipedia.org/.