asafaya

/

bert-base-arabic

like 35

Hey guys, you did a great job in making this one. However I am trying to use bert to train a language model for Amharic too. So I would like to get heads up on how you managed to train it. Like what type of data did you use? Did you use vectorized tokens like as in Word2vec? Bit confused. It is a personal project.

Thanks :P