I'm implementing https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt for translations as it has shown promising results but I wanted to see if there was a way to translate specific parts differently with a list of key terms. For example if it translates "Space Odyssey: 2000" as "odessia espasial de 2000" I would prefer not to translate that part, or a different way. Is there a technique with masking I could implement or is it a matter of more fine tuning with specific sentences? I have a long list of pre translated domain specific terms that are not common or are standard translations I require.

Trying to text replace post translation would require to have all possible variations of the "incorrect" translation is a huge undertaking with thousands of specific terms.


Your Answer

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

Browse other questions tagged or ask your own question.