metadata

dataset_info:
  features:
    - name: text
      dtype: string
    - name: sentences
      sequence: string
    - name: parses
      sequence: string
  splits:
    - name: train
      num_bytes: 713719296
      num_examples: 769764
  download_size: 338028239
  dataset_size: 713719296
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
language:
  - en
tags:
  - wikipedia
  - wiki
size_categories:
  - 100K<n<1M

Dataset Card

This dataset is a processed version of rahular/simple-wikipedia, which is a dump of articles from Simple English Wikipedia.

In addition to the raw texts in the text column, this dataset also provides two additional columns:

sentences: A list of sentences in text, produced by a spaCy sentence tokenizer
parses: A list of constituency parse strings, one per sentence in sentences, generated by the Berkeley neural parser