- Clone the repo.
- Change into the repo directory and run
make init(or justmake) to download the datasets. - Optionally run
make optionalto obtain additional data resources (eg. for subword-units). Requires authentication. - Run
make runto run an experiment with default settings.
kerasgensimnumpy(implied by the previous anyway)
subword-nmt