Introduction

This is one of the capstone projects of the Natural Language Processing Nanodegree from Udacity, where I built an deep neural network that functions as part of an end-to-end machine translation pipeline. The model accepts English text as input and return the French translation.

In this notebook, you will find several different RNN models that I have built and experiment different architecture designs. The final mode is an Encoder-Decoder model that is incorporate with word embeddings.

Setup

For this project, it is advisable to use GPU acceleration to train the model.

Libraries needed

• Python 3
• NumPy
• TensorFlow 1.x
• Keras 2.x