An encoder-decoder model includes an [[encoder]] and [[decoder]]. Common use cases include [[machine translation]], text summarization, and question answering. The encoder and decoder layers in a [[transformer]]-based architecture are linked with the [[cross-attention]] layer.