Tokenizer/word breaker: BPE (GPT, Roberta); WordPiece (Bert, Electra); SentencePiece (AlBert, T5)

Word breaker: BPE (GPT, Roberta); WordPiece (Bert, Electra); SentencePiece (AlBert, T5)

Guess you like

Origin blog.csdn.net/u013250861/article/details/131254531