BabyLM-community/babylm-fas

BabyLM Dataset

Dataset Description

This dataset is part of the BabyLM multilingual collection.
More information at: babylm.github.io/babybabellm

Dataset Summary

• Language: fas
• Script: Arab
• Tier: 100M
• Byte Premium Factor: 1.597326
• Size (MB): 867.30
• Expected Size (MB): 867.35
• Number of Documents: 217,776
• Total Tokens: 98,506,081
• Tokenizer: separate by whitespace

Tokens Per Category

• child-books: 67,165 tokens
• educational: 94,320,928 tokens
• padding-opensubtitles: 4,117,988 tokens

Tokens Per Group

• Transcription: 0 tokens
• Education: 94,320,928 tokens
• Books, Wiki, News: 67,165 tokens
• Subtitles: 0 tokens
• Padding: 4,117,988 tokens

Data Fields

• text: The document text
• doc-id: Unique identifier for the document
• category: Type of content (e.g., child-directed-speech, educational, etc.)
• data-source: Original source of the data
• script: Writing system used (ISO 15924)
• age-estimate: Target age or age range
• license: Data license
• misc: Additional metadata (JSON string)
• num-tokens: Number of tokens per item (based on white-space split)
• language: Language code (ISO 639-3)

Licensing Information

Please see license in individual documents

Data Sources & Attribution

Books

• Ririro Story Collection
  • source: https://ririro.com/
  • description: Collection of children storybooks
• GlotStoryBooks Collection
  • source: https://arxiv.org/abs/2310.16248
  • description: Collection of children storybooks

Educational

• FineWeb2-HQ
  • source: https://arxiv.org/abs/2502.10361
  • description: A high-quality, multilingual dataset built on top of FineWeb2 (Penedo et al., 2025) filtered for educational content

Data Curators

• Negar Foroutan
• Pouya Sadeghi

Comments

None