Professional Terms and Phrases Finetuning
When I train with professional terms and phrases,like "Journey to the West--西游记", the semantics of the original model are affected, even can not translate zh to zh . How can i solve it?
So far, the functionality of working with professional terms is not provided. This is planned to be implemented in future versions.
Thanks. I am a beginner in NLP and now I need to train on professional terms and phrases. I have tried fine-tuning several models before, but the results are not good. Can we exchange ideas on this topic?
So far, the functionality of working with professional terms is not provided. This is planned to be implemented in future versions.
Yes, it is very interesting. It seems that the main problem here is data collection. But the examples of submitting information about terms to the model are also interesting. Perhaps there are some best practices here.
Can we add a contact, like Facebook、Ins?
telegram @utrobinmv