LoneWolfgang's picture
Update README.md
b85b82f verified
metadata
dataset_info:
  features:
    - name: name
      dtype: string
    - name: homepage
      dtype: string
    - name: domain
      dtype: string
    - name: subdomain
      dtype: string
    - name: quality
      dtype: int64
    - name: notes
      dtype: string
    - name: ids
      sequence: string
    - name: id_count
      dtype: int64
  splits:
    - name: train
      num_bytes: 586272
      num_examples: 29
  download_size: 447242
  dataset_size: 586272
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

This a collection of IDs that potentially match to high-quality educational material from the Japanese segment of FineWeb.

Total websites: 29 Total IDs: 11,346

My goal is to pick out IDs from 50 websites. If you are following this, please check back later.

The columns are:

  • name (str): Shorthand indentifier of domain where IDs were found.
  • homepage (str): Full URL of the domain where IDs were found.
  • domain (str): Coarse-grain topical classification.
  • subdomain (str): Fine-grain topical classification.
  • quality (int): Educational quality ranked 1-5, with 5 being the best quality.
  • notes (str): Explains the content of the website and any shortcomings.
  • ids (List(str)): List of FineWeb IDs that were selected from the website.
  • id_count (int): Number of IDs that were selected from the website.