levuloihust's picture
Create README.md
45a7d03
metadata
language:
  - vi
  - en

Description

The file vien-corpus.txt has 7957186 lines, mixing Vietnamese and English texts. This file was used as a corpus to train a tokenizer for both Vietnamese and English.

Details

How was this corpus created?

  1. English text is from Wikipedia
  2. Vietnamese text is from 2 sources:
  • Crawled data from news websites
  • Oscar dataset