Creating corpuses

A corpus contains sentence pairs that are used to adapt machine translation models.‍

To create a corpus, go to SOURCE and under the CORPUS tab, click CREATE CORPUS.

Enter the name for the corpus and optionally upload a CSV file, then click CREATE

🚧

Requirement

Make sure the CSV file is formatted properly.

There are two ways to add sentence pairs to a corpus:

1. Import from a CSV file

  • Click IMPORT PAIRS
  • See the note above regarding file format
  • The corpus will show a preview of the sentence pairs contained in the CSV file
  • When the sentence pairs are finalized, click IMPORT (X) PAIRS

2. Create sentence pairs manually

  • Add matching pairs in the source and target language
  • Click ADD PAIRS
  • Enter the desired phrase in the source language and target language
  • When the phrases are finalized, click ADD (X)