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