Training a paraphrase model in other language
#3
by thebluetick - opened
I have tested the model and am very impressed with its inference. Currently, I am also training a similar paraphrase model for Vietnamese, but the results are often grammatically incorrect and do not retain the entities in the sentence. Any suggestions to keep entities in a sentence?
Currently there is no PEGASUS model for Vietnamese language, so I use T5 instead. Is that part of the problem?