plan on multilingual variant?
#2
by ahxxm - opened
from tokenizer.json and section Training in blog post, it seems an English-only(the same naming recipe as origin BERT I guess)
I come to the community section to ask the same question, any ideas on the multilingual part ?
I join everyone to beg for a wondeful ModernXLM-RoBERTa-Large! :)
I came for the same - I need one that works well across as many languages as possible