README.md

metadata

annotations_creators:
  - no-annotation
language_creators:
  - found
language:
  - en
license:
  - mit
multilinguality:
  - monolingual
pretty_name: enwik8
size_categories:
  - 10K<n<100K
source_datasets:
  - original
task_categories:
  - fill-mask
  - text-generation
task_ids:
  - language-modeling
  - masked-language-modeling
dataset_info:
  - config_name: enwik8
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 104299244
        num_examples: 1128024
    download_size: 36445475
    dataset_size: 102383126
  - config_name: enwik8-raw
    features:
      - name: text
        dtype: string
    splits:
      - name: train
        num_bytes: 100000008
        num_examples: 1
    download_size: 36445475
    dataset_size: 100000008

Dataset Card for enwik8

Table of Contents

• Dataset Description
  • Dataset Summary
  • Supported Tasks
  • Languages
• Dataset Structure
  • Data Instances
  • Data Fields
  • Data Splits
• Dataset Creation
  • Curation Rationale
  • Source Data
  • Annotations
  • Personal and Sensitive Information
• Considerations for Using the Data
  • Social Impact of Dataset
  • Discussion of Biases
  • Other Known Limitations
• Additional Information
  • Dataset Curators
  • Licensing Information
  • Citation Information

Dataset Description

• Homepage: http://mattmahoney.net/dc/textdata.html
• Repository: [Needs More Information]
• Paper: [Needs More Information]
• Leaderboard: https://paperswithcode.com/sota/language-modelling-on-enwiki8
• Point of Contact: [Needs More Information]
• Size of downloaded dataset files: 36.45 MB
• Size of the generated dataset: 102.38 MB
• Total amount of disk used: 138.83 MB

Dataset Summary

The enwik8 dataset is the first 100,000,000 (100M) bytes of the English Wikipedia XML dump on Mar. 3, 2006 and is typically used to measure a model's ability to compress data.

Supported Tasks and Leaderboards

A leaderboard for byte-level causal language modelling can be found on paperswithcode

Languages

en

Dataset Structure

Data Instances

• Size of downloaded dataset files: 36.45 MB
• Size of the generated dataset: 102.38 MB
• Total amount of disk used: 138.83 MB

{
   "text": "In [[Denmark]], the [[Freetown Christiania]] was created in downtown [[Copenhagen]]....",
}

Data Fields

The data fields are the same among all sets.

enwik8

• text: a string feature.

enwik8-raw

• text: a string feature.

Data Splits

dataset	train
enwik8	1128024
enwik8- raw	1

Dataset Creation

Curation Rationale

[Needs More Information]

Source Data

Initial Data Collection and Normalization

The data is just English Wikipedia XML dump on Mar. 3, 2006 split by line for enwik8 and not split by line for enwik8-raw.

Who are the source language producers?

[Needs More Information]

Annotations

Annotation process

[Needs More Information]

Who are the annotators?

[Needs More Information]

Personal and Sensitive Information

[Needs More Information]

Considerations for Using the Data

Social Impact of Dataset

[Needs More Information]

Discussion of Biases

[Needs More Information]

Other Known Limitations

[Needs More Information]

Additional Information

Dataset Curators

[Needs More Information]

Licensing Information

[Needs More Information]

Citation Information

Dataset is not part of a publication, and can therefore not be cited.

Contributions

Thanks to @HallerPatrick for adding this dataset and @mtanghu for updating it.