0
$\begingroup$

I want to build a machine translation system from English to Georgian. Georgian is a language similar (and simpler) to the Russian language. its syntax looks like base + suffix, only suffix changes, most of the time base is frozen, to describe the time the only suffix is changed. Unfortunately, I couldn't find a morphological analyser for the Georgian language, so could you link or provide useful resources to help me to build one? or can you give me some suggestions?

$\endgroup$
3
  • $\begingroup$ Do you mean a tokeniser or a morphological analyser? $\endgroup$ – Oliver Mason May 26 at 8:12
  • $\begingroup$ i mean tokenizer, i have text( georgian, i will mention english for everybody to understand), "i am dato nefaridze, and living in Georgia", i think it should split these text into words ['i', 'am' 'dato', 'nefaridze' 'and' 'live' 'ing' 'in' 'Georgia']. it should put apart living word $\endgroup$ – dato nefaridze May 26 at 17:47
  • 1
    $\begingroup$ That's not a tokeniser. A tokeniser is mostly language-agnostic and only splits a stream of characters into words and non-words. You're looking for a morphological analyser, which can identify component parts of words, and therefore relies on knowledge about the structure of the language. $\endgroup$ – Oliver Mason May 27 at 8:46
1
$\begingroup$

Huggingface has a Helsinki-NLP/opus-mt-ka-en repository with a Georgian (Ka) to English (en) model. A tokenizer_config.json is available

$\endgroup$
1
  • $\begingroup$ i want to build my own translation system. i think it is already trained model for translation, i need tokenizer not whole model. Also i want english to georgian, not vice versa $\endgroup$ – dato nefaridze May 26 at 17:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.