Deploy a Spark model from Flow Editor

When you work with a Spark machine learning flow in the Flow Editor interface, you can deploy the model directly by right-clicking a node and selecting the Deploy option. Once you've saved the model from the Flow Editor, it appears in your list of models in the Watson Studio project view and can be deployed from there as well.

Note: The deploy context-menu item is present for all nodes except source and model build nodes. A machine learning flow can only be deployed if it contains a trained model and will error if no model object is present.

Flow Editor deployment steps

When you use the Flow Editor to create a machine learning flow, you must first save the machine learning flow as a model before you can deploy it.

  1. First, save the machine learning flow.
    1. Right-click the model nugget node.
    2. Click Save as a model.
    3. In the Model Name box type a name.
    4. Click Save.
  2. Then, deploy this model.
    1. Return to the project.
    2. From the Assets tab, in the Models section, click the model
    3. Click Add Deployment.
    4. In the Deployment Type box, select Online.
    5. In the Name box, type a name.
    6. Click Deploy.

Next steps

For more information about continuous learning and model evaluation, see Continuous learning and model evaluation