An encoder converts an input such as text, an image, audio, or structured data into an internal numerical representation. In an encoder–decoder system, that representation is passed to a decoder that produces a translation, summary, reconstruction, or other output. Encoder-only transformers such as BERT are often used for classification, retrieval, and feature extraction because they build context-aware representations of the full input. Autoencoders learn encodings by reconstructing their original data. The usefulness of an encoder depends on its training objective and dataset; a compact vector may retain information needed for one task while discarding detail required for another.
Google Cloud and Figma have launched an expanded partnership to embed Google’s generative AI models – including Gemini 2.5 Flash and Imagen 4 – directly into Figma’s design tools, reducing latency and accelerating creative workflows.