Deploying ML models in Snowpark

In the preceding chapter, we learned about how to develop ML models. Now that the models are ready, we must deploy them into Snowpark. To make it easier for developers to deploy the models, the Snowpark ML library consists of functions that encompass the introduction of a new development interface and additional functionalities aimed at securely facilitating the deployment of both features and models. Snowpark MLOps seamlessly complements the Snowpark ML Development API by offering advanced model management capabilities and integrated deployment functionalities within the Snowflake ecosystem. In the following subsections, we will explore the model registry and deploy the model for inference to obtain predictions.

Snowpark ML model registry

A model registry is a centralized repository that enables model developers to organize, share, and publish ML models efficiently. It streamlines collaboration among teams and stakeholders, facilitating the collaborative...