Deploying AI assets programmatically

You can deploy your AI assets in Watson Machine Learning programmatically by using watsonx.ai Python client library, Watson Machine Learning API, and IBM Cloud Pak for Data Command Line Interface (IBM cpdctl).

Deploying assets with watsonx.ai Python client library

Use the watsonx.ai Python client library to work with traditional machine learning and generative AI assets to train, store, and deploy your models. You can score your models by using the APIs and integrate them with your application deployment. To learn more about the installation, setup, and usage, see ibm-watsonx-ai library documentation.

Note:

The Watson Machine Learning Python client library (ibm-watson-machine-learning) is now part of the expanded watsonx.ai Python client library (ibm-watsonx-ai). The Watson Machine Learning library will persist but will not be updated with new features.

Sample notebooks

You can refer to sample notebooks that demonstrate machine learning features and techniques.

For more information, see Watson Machine Learning Python client samples and examples.

Deploying assets with APIs

You can use the watsonx.ai API, Watson Machine Learning API, and Watson Data API to build models and deploy them for use in applications. Choose from tools that fully automate the training process for rapid prototyping to tools that give you complete control to create a model that matches your needs. You can access the Watson Machine Learning REST API from the Python client library.

Use the Watson Machine Learning API to deploy AutoAI experiments, models, trainings, and pipelines and more.

For more information, see Watson Machine Learning API documentation.

Use the Watson Data API to deploy and manage data-related assets such as assets, catalogs, environments, hardware specifications, software specifications, runtime definitions, and more.

For more information, see Watson Data API documentation.

Deploying assets with IBM Cloud Pak for Data Command Line Interface

Use the IBM Cloud Pak for Data Command Line Interface (cpdctl) to manage the configuration settings and automate the lifecycle of AI assets. You can create or delete projects, change project hardware and software specifications, update package extensions, and manage access. You can also prepare and manage data assets and connections, create experiments, deploy models, and create pipelines.

For more information, see Managing AI lifecycle with cpdctl.

Authenticating for programmatic access

You must authenticate to use Watson Machine Learning securely by using a token or credentials. The preferred method to authenticate with a token is by using ZenApiKey. If you are not storing your notebook within Watson Studio, you can bypass retrieving a token and authenticate with your Cloud Pak for Data credentials.

For more information, see Authenticating for programmatic access.

Learn more

Parent topic: Deploying AI assets