multv-nmt

data

put data under a folder named data/

make config files for training

Change the directory name, random seed, language code as desired in make-cfg.sh, then run ./make-cfg.sh The config files are written under the folder scripts/

Decode

Make appropriate changes to make-trans.sh, then run python make-trans.py The translation config files will be written to the corresponding directory under scripts/ This will only create translation config for the output folder that does not contain a translation already.