how to train in another languages?

[YusufCandraArif]

Thank you for this amazing project.
Is it possible to be trained in another language, for example Mandarin? Thank you.

[oliverguhr]

Hi @YusufCandraArif,
basically you would need to create a dataset for your language in the same format, as the SEPP NLG task that we build this model for. You can find the details here.

After that you need to add a config entry for your task in this file. After this you can start the training via:

python model_test_suite.py 2

And you are done :) Typically, some issues appear on the way. If you have any questions drop me a line or open a issue.

[oliverguhr]

Thanks to @jordimas we have a how to guide on this now:
https://github.com/oliverguhr/fullstop-deep-punctuation-prediction/blob/main/other_languages/readme.md

[YusufCandraArif]

Thank you Sir @oliverguhr