I O - All #NLP taks need a fast tokenizer. I thought we are not going to have tokenizer faster than O(nm). But this tokenizer work with O(n) #algorithm https://t.co/cOHeoANIsq

4:39am PDT - October 10th, 2021

All #NLP taks need a fast tokenizer. I thought we are not going to have tokenizer faster than O(nm). But this tokenizer work with O(n) #algorithm https://t.co/cOHeoANIsq

0 retweets 0 likes -