I want to upgrade the en_hu (English - Hungarian) .argosmodel translate pack, but I can’t figure out, how to get the original files (source and target) from the trained language package.
Is there any method to get the source files? Or is there any online source for the .argosdata files?
I made at least 8-10 version from zero based on Opus corpus, but to upgrade the existing version seems to be much better option.
Thank you in advance.
I have links to data for some languages in argos-train/data-index.json but I don’t currently have any packaged for Hungarian. I get most of the data from Opus so you’ll need to get it directly from there.
I’m also happy to accept pull requests adding new data packages, I’d like to expand the data index which Argos Train can use in an automated way.