To run evaluation of existing models:
python3 -W ignore loaded.py language-option
Warnings are ignored for BLEU score calculations
loaded.py should be modified for whatever evaluation functions you want to use
If you just want to play around I recommend using evaluateRandomly or evaluateFromInput :)
To train new models:
python3 seq2seq.py language-option
language-option: e2s is English to Spanish, s2e is Spanish to English
Must use pytorch v 1.0.0