Hello! Bert and many other pre-trained language models use the Wordpiece algorithm so stemming is not necessary.

May be useful

https://analyticsindiamag.com/a-guide-to-text-preprocessing-using-bert/

Hi! Preprocessing is not needed when using pre-trained language representation models like BERT. In particular, it uses all of the information in a sentence, even punctuation and stop-words, from a wide range of perspectives by leveraging a multi-head self attention mechanism.

You can find some well explained examples + GitHub code here: https://riccardo-cantini.netlify.app/post/

Hope it helps.

Cheers

Actually when I used preprocessing steps I saw worse results than without it. But I want to make sure.

Hello, I agree with you, Mirzaei. After stemming, BERT's performance of text classification decreased.

hello mahsa, i want to see your project if u agree, cause i need to see the preprocess step