Skip to content

Model Registry Guides#

Hopsworks Model Registry is a centralized repository, within an organization, to manage machine learning models. A model is the product of training a machine learning algorithm with training data.

This section provides guides for creating models and publish them to the Model Registry to make them available for download for batch predictions, or deployed to serve realtime applications.

Exporting a model#

Follow these framework-specific guides to export a Model to the Model Registry.

  • TensorFlow


    Export a TensorFlow or Keras model.

    Export guide

  • Torch


    Export a PyTorch model.

    Export guide

  • Scikit-learn


    Export a scikit-learn model.

    Export guide

  • LLM


    Export a large language model.

    Export guide

  • Triton


    Export a Triton model directory for NVIDIA Triton Inference Server.

    Export guide

  • Other Python frameworks


    Export any other Python model, such as XGBoost or LightGBM.

    Export guide

Importing a model from HuggingFace#

You can also import a model directly from the HuggingFace Hub through the Hopsworks UI. The download runs server-side and the model is registered automatically. See Import from HuggingFace.

Model Schema#

A Model schema describes the input and outputs for a model. It provides a functional description of the model which makes it simpler to get started working with it. For example if the model inputs a tensor, the model schema can define the shape and data type of the tensor.

Input Example#

An Input example provides an instance of a valid model input. Input examples are stored with the model as separate artifacts.