This project is partly derived from the official tutorial, an example which trains a sequence-to-sequence (seq2seq) model for Spanish to English translation.
Please follow the official guide to install Tensorflow. In addition, you need to install scikit-learn.
prepare_dataset.py downloads a parallel corpus with some preprocessing.
The resultant files will consume about 8.1MB.
$ python prepare_dataset.py <directory-to-save-dataset>
train.py starts training a seq2seq model on the dataset.
$ python train.py --dataset <directory-to-save-dataset> --device 0