Importing an ONNX model

You can train an ONNX model on your local distributed platform and then import it into your ML for IBM z/OS® for deployment and management.

Before you begin

Make sure that you have completed the following installation and configuration tasks:

If you run your ML for IBM z/OS on z/OS V2.5 or higher and plan to use the micro-batching inference and on-chip AI acceleration capabilities for online scoring of your ONNX models or snapML models, apply the PTFs that install the IBM® Z Deep Neural Network Library (zDNN), as described in Installing prerequisite hardware and software for ML for IBM z/OS.
Provision a required z/OS Container Extensions instance as described in Installing prerequisite hardware and software for ML for IBM z/OS.
Configure an ONNX compiler service for your MLz environment as described in Configuring the ONNX compiler service for your ML for IBM z/OS.

Sign in to the ML for IBM z/OS user interface at https://<yourMLzUI-URL> with your username and password, where <yourWMLzUI-URL> is the host IP address and port number for your MLz UI.

See Configuring MLz for more information about the host IP address and port number of your MLz.
Go to the Models tab and click Import model.
On the Import model page, click the From file tab.
Specify a name for the new model and select ONNX from the model type menu.
Upload your model source file in .onnx format.

You can simply drag and drop or browse for the source file into the Upload box.
Click Import model to import the new model.
Verify that the imported model shows up on the Models page.

Tip: Do not use the new model until it is successfully compiled and its status on the Models page is valid.