On a slight tangent, are there any fully-trained ready-to-work state-of-the-art open source distributions of translation systems available?

I've never been able to find one, but maybe I just haven't looked hard enough.

Both Google and Facebook have released pre-trained models in a few languages [0, 1].

[0] https://github.com/facebookresearch/fairseq

[1] https://google.github.io/seq2seq/nmt/