README.md

metadata

dataset_info:
  features:
    - name: url
      dtype: string
    - name: html_name
      dtype: string
    - name: skipped
      dtype: string
    - name: markdown
      dtype: string
    - name: text
      dtype: string
    - name: batch
      dtype: string
    - name: doc_id
      dtype: string
    - name: split_markdown
      sequence: string
    - name: slide_markdown
      sequence: string
  splits:
    - name: train
      num_bytes: 544983664
      num_examples: 223
  download_size: 40899249
  dataset_size: 544983664
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
pretty_name: Basic Evaluation dataset for IR

Preprocess text

method: window_slide step: 50 chunk_size: 1500