1. Learn
    2. /
  2. Courses
    3. /
  3. Recurrent Neural Networks (RNN) for Language Modeling in Python

Exercise

Translate Portuguese to English

This is the last exercise of the course, congratulations on getting here!

You will learn how to use NMT models for making translations.

A model that encodes Portuguese small phrases and decodes them into English small phrases was pre-trained and is loaded in the model variable.

Also, the function predict_one() is already loaded, use help() for details and the dataset is available on the test (raw text) and X_test (tokenized) variables.

You will define a function to translate a list of sentences. In the parameters, sentences is a list of phrases to be translated, index_to_word is a dict containing numerical indexes as keys and words as values for the English language, loaded in the en_index_to_word variable.

The model summary has been printed for your consideration.

Instructions

100 XP
  • Loop over the enumerated iterator of the phrases.
  • Use the pre-loaded function predict_one() to translate one phrase.
  • Print the translation result.
  • Call the defined function to translate the initial 10 phrases of the X_test variable.