Train only monolingual files in the hub
I only have monolingual tmx files and i want to train them can i do that on the hub ?
Do you have any parallel sentences or just these monolingual ones? We don't tune on just monolingual data for now. You need to have atleast dev and test sets and some sort of a dictionary to build a model that can still use the Microsoft models.
Hello @vishalc i've triedto upload those monolingual files but i can't start the training and i get the message as there not enough parallel documents to start the training
We officially support only bilingual tmx files. Did you try to upload your files on the Hub? What was the outcome?