TensorFlow Serving is a production-serving system for deploying trained machine learning models. It provides infrastructure for exposing models to applications and handling inference requests reliably.
A platform to deploy trained models in production.