Deploying Models

There are several ways to deploy your model into your Modzy instance. As the model developer, you can decide how you package and import your model based on the framework or language you are working with.

Deploy Now

No-Code Model Import Options:

• AWS Sagemaker
• Azure ML
• MLflow

Deploy an OMI-Compliant Container

• Import pre-built Docker container you have already built that meets the Open Model Interface Specification

Automatically Build and Deploy Model Container

• PyTorch
• Scikit-learn
• XGBoost
• LightGBM
• Fastai
• MXNet
• ONNX
• PMML
• General Chassis Template - if you do not see your model framework in the above list, it is likely supported! Follow this general template to automatically build and containerize your model (Note: Tensorflow/Keras is the only known framework currently not supported by Chassis)

Manually Create OMI-Compliant Container

• Python Model
• R Model

Set Up a CI/CD Pipeline

• Github Actions