Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Abbreviations Issue #64

Open
Khalid-kamal opened this issue Jan 28, 2023 · 5 comments
Open

Abbreviations Issue #64

Khalid-kamal opened this issue Jan 28, 2023 · 5 comments

Comments

@Khalid-kamal
Copy link

Dear Tommi,
When I add a translation for an abbreviation, all other abbreviations takes the same translation
for example:
image
The same Translation for different abbreviations

@TommiNieminen
Copy link
Collaborator

How did you add the translation, do you mean you fine-tuned the model with data that contained the translation for the abbreviation? NMT models can be a bit fuzzy at times, especially in cases such as these where there are many slightly different words in the same sentence. The translations for the abbreviations may be different, if you try them out in different contexts.

@Khalid-kamal
Copy link
Author

So, why does it have the same meaning for different abbreviations?

@Khalid-kamal
Copy link
Author

For abbreviations that the tool cannot recognize, it leaves them as is KMo for instance. It should leave the abbreviations that do not have equivalents as in source, not translating them to anything even if it has no equivalent.

@Khalid-kamal
Copy link
Author

Is there a new version of the tool?

@TommiNieminen
Copy link
Collaborator

The NMT models that are used in OPUS-CAT are difficult to control, they tend to improvize a lot, so you end up with strange behavior with words and abbreviations that have not occurred in the training data used to create the model. If the mistranslation occurs regularly, you can create an edit rule to correct it automatically, but that requires some manual work.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants