metadata

language:
  - en
  - es
  - fi
  - fr
  - pt
  - nl
  - de
  - sv
  - it
license: odc-by
size_categories:
  - 1K<n<10K
task_categories:
  - text-generation
  - feature-extraction
dataset_info:
  - config_name: default
    features:
      - name: filename
        dtype: string
      - name: text
        dtype: string
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 2190406903
        num_examples: 4999
    download_size: 1310735987
    dataset_size: 2190406903
  - config_name: en-clean
    features:
      - name: filename
        dtype: string
      - name: text
        dtype: string
      - name: lang
        dtype: string
      - name: score
        dtype: float64
    splits:
      - name: train
        num_bytes: 1948279468.5718827
        num_examples: 4371
    download_size: 1167660575
    dataset_size: 1948279468.5718827
  - config_name: english
    features:
      - name: filename
        dtype: string
      - name: text
        dtype: string
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 1929258170.415883
        num_examples: 4403
    download_size: 1172514726
    dataset_size: 1929258170.415883
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
  - config_name: en-clean
    data_files:
      - split: train
        path: en-clean/train-*
  - config_name: english
    data_files:
      - split: train
        path: english/train-*
tags:
  - gutenberg
  - epub
  - marker

BEE-spoke-data/napierone-epub-raw

NapierOne EPUB files converted with marker. Seems to contain mostly books from Project Gutenberg.

detected languages

via fasttext-langdetect

{'ca': 1,
 'cy': 1,
 'da': 6,
 'de': 105,
 'en': 4403,
 'eo': 2,
 'es': 61,
 'fi': 76,
 'fr': 189,
 'he': 1,
 'hu': 5,
 'is': 1,
 'it': 40,
 'la': 6,
 'nl': 41,
 'pl': 4,
 'pt': 38,
 'sv': 10,
 'tl': 9}