How_to_make_a_language_translator
This is the code for "How to Make a Language Translator - Intro to Deep Learning #11' by Siraj Raval on YouTube
Coding Challenge - Due Date Thursday March 30th at 12 PM PST
Make your own language translator using a deep learning architecture of your choice in Tensorflow. It doesn't have to use Neural Machine Translation specifically. Here are some example. This exercise will help you appreciate the full power of statistical translation techniques.
Overview
This is the code for this video on Youtube by Siraj Raval as part of the Deep Learning Nanodegree course with Udacity. This code implemnents Neural Machine Translation which is what Google Translate uses to translate Languages.
Dependencies
- tensorflow
- nltk
- six
Install missing dependencies with pip
Usage
To train model on data and test it to compute the BLEU score run this:
python translate.py source_language target_language
(i.e. python translate.py fr en for fr->en translation)
Training results are shown in plot_training_process.ipynb.
Testing results and findings are discussed in the paper.
Credits
The credits for this code go to fanshi118. I've merely created a wrapper to get people started.