README.md

metadata

license: cc-by-nc-4.0
language:
  - en
  - ar
size_categories:
  - 100M<n<1B
task_categories:
  - text-generation

DeepLatent Pretraining Data

Combined English and Arabic pretraining data for DeepLatent language models.

Dataset Statistics

Language	Rows
English	40,182,085
Arabic	91,363,714
Total	131,545,799

Schema

• text (string): The document text
• language (string): Language code ('en' or 'ar')

Data Sources

• English: FineWeb-Edu sample
• Arabic: Deduplicated Arabic web corpus

Usage

from datasets import load_dataset

# Load full dataset
dataset = load_dataset("almaghrabima/deeplatent-pretraining-data")

# Load specific language
english = dataset.filter(lambda x: x['language'] == 'en')
arabic = dataset.filter(lambda x: x['language'] == 'ar')

License

cc-by-nc-4.0