WordPiece

E1339930 UNEXPLORED

WordPiece is a subword tokenization algorithm that breaks text into smaller units to efficiently handle rare and unknown words in language models.

All labels observed (1)

Label Occurrences
WordPiece canonical 1

How this entity was disambiguated

Referenced by (1)

Full triples — surface form annotated when it differs from this entity's canonical label.

BERT usesTokenization WordPiece