Back to the Index

Precomputed Features

What is a precomputed feature?

A precomputed feature is a feature that has been created by a feature pipeline and is stored in a feature store. Training and inference pipelines can read precomputed features and use them directly for training or inference. Precomputed features are typically stored unencoded if they are intended for reuse across multiple models, but they can also be stored encoded if an online model has low latency SLAs and encoding the feature at runtime is too expensive.

Does this content look outdated? If you are interested in helping us maintain this, feel free to contact us.

© Hopsworks 2024. All rights reserved. Various trademarks held by their respective owners.

Privacy Policy
Cookie Policy
Terms and Conditions