Loading…
Glossary term
The part of a model that turns input into an internal representation.
An encoder reads an input and turns it into a compact internal representation, often a list of vectors capturing its meaning.
Encoder-only models, such as BERT, are good at understanding tasks like classification and search. Encoder-decoder models, used for translation, hand the representation to a decoder.
Encoders are the 'reading' half of many models and the basis of embedding models used in search.