Translating your dataset

Nádia Dias Updated by Nádia Dias

Translating your dataset

One of the main features of Weni is the ability to translate your dataset into various languages, allowing your intelligence to start predicting and understanding inputs from all over the world.

In this article, you will learn how to manage a new language in your intelligence and track the progress of existing translations.

Translating Phrases

The translation process is simple and straightforward. First, you will need to access the Translation tab.

Select the base language that needs to be translated.

After that, the platform will display a list of all the phrases that have not yet been translated into these languages in the current dataset. You can filter the phrases by intent and entities, if needed.

Now, select a phrase to translate and click on 'Translate sentences'.

These phrases are translated and will be removed from the list of sentences not translated.

Translate the other phrases in your dataset, and your intelligence will be able to recognize this new language!

Important Notes

To train a new language, you need to have at least 2 phrases translated for each intent and entity in your original dataset, or you will receive the alert below. You will see this error inside the 'Training' tab.

In this case, the translation to Brazilian Portuguese (pt_br) contains only one phrase for each intent in the intelligence. To fix this, translate more phrases so that the training can be activated normally.

Deleting a Translation

In the 'Translation' tab, you can also delete translations from your dataset. There is a list of translated phrases.

To delete, click 'Delete' in front of the desired phrase.

Now, this phrase will appear in the Sentences not translated tab and can be translated again.

How did we do?

Interactions

Versioning

Contact