Results, reports and models of the project "Kama: usable Estonian machine translation (kasutatav eesti masintõlge)"
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Kama: Usable Estonian Machine Translation (kasutatav masintõlge) is a project (2015-2017) under the National Programme for Estonian Language Technology. Its focus is improving machine translation from/into Estonian and making sure that its output is usable in practice for specific text domains, for example for post-editing.

Here we publicly distribute the latest SMT and NMT models that we have developed, as well as their analysis and test set translations.


We are working on word2word and chr2chr neural MT approaches. You can find the model files, translation quality estimations and references to used software in the nmt folder.


Our previous baseline was phrase-based SMT, and included a number of improvements on top of the baseline approach. In the smt folder you will find the models, analysis of translation quality and other info.