Quick start: Build and deploy a machine learning model in a Jupyter notebook

You need a project to store the data and the AutoAI experiment. You can use an existing project or create a project.

1. From the Navigation Menu , choose Projects > All projects.

2. Open an existing project. If you want to use a new project:

   1. Click New project.

   2. Select Create an empty project.

   3. Enter a name and optional description for the project.

   4. Click Create.

For more information or to watch a video, see Creating a project.

You will use a sample notebook in this tutorial. Follow these steps to add the sample notebook to your project:

1. From the Assets tab in the project, click New asset > Work with data and models in Python or R notebooks.

2. Select the URL page.

3. Paste the following link in the Notebook URL field:

   https://github.com/IBM/watsonx-ai-samples/blob/master/cpd4.8/notebooks/python_sdk/deployments/scikit-learn/Use%20scikit-learn%20to%20recognize%20hand-written%20digits.ipynb
   

4. Type the notebook name and description (optional).

5. Select a runtime environment for this notebook.

6. Click Create. Wait for the notebook editor to load.

7. From the menu, click Kernel > Restart Kernel and Clear Outputs of All Cells, then confirm by clicking Restart to clear the output from the last saved run.

The first section in the notebook sets up the environment by specifying your Cloud Pak for Data credentials and Watson Machine Learning service instance location. Follow these steps to set up the environment in your notebook:

1. Scroll to the Set up the environment section.

2. Obtain your API key. If you do not have an API key, follow these steps:

   1. Open another tab with your Cloud Pak for Data deployment.

   2. Click your Profile icon.

   3. Click Profile and settings.

   4. Click API Key > Generate new key.

   5. Copy your API key and close the tab.

3. The platform URL is the hostname or URL that you use to access your Cloud Pak for Data deployment. Ask your administrator if you don't have this information. The administrator can obtain the platform URL from the Red Hat OpenShift console (Red Hat OpenShift Container Platform -> Projects -> cpd-instance -> Routes).

4. Paste your username, api_key, and platform url into cell 1.

5. Click the Run icon to run your code in cells 1 and 2.

6. Run cell 3 to install the ibm-watson-machine-learning package.

7. Run cell 4 to import the API client and create the API client instance using your credentials.

8. Run the cell with the code client.spaces.list(limit=10) to see a list of all existing deployment spaces. If you do not have a deployment space, then follow these steps:

   1. Open another tab with your Cloud Pak for Data deployment.

   2. From the Navigation Menu , click Deployments.

   3. Click New deployment space.

   4. Add a name and optional description for the deployment.

   5. Click Create, then View new space.

   6. Click the Manage tab.

   7. Copy the Space GUID and close the tab, this value will be your space_id.

9. Copy and paste the appropriate deployment space ID into the cell with the code space_id = 'PASTE YOUR SPACE ID HERE', then run that cell and the cell with the code client.set.default_space(space_id) to set the default space.

Now that all of the environment variables are set up, you can run the rest of the cells in the notebook. Follow these steps to read through the comments, run the cells, and review the output:

1. Run the cells in the Explore data section.

2. Run the cells in the Create a scikit-learn model section to.

   1. Prepare the data by splitting it into three data sets (train, test, and score).

   2. Create the pipeline.

   3. Train the model.

   4. Evaluate the model using the test data.

3. Run the cells in the Persist locally created scikit-learn model section to publish the model, get model details, and get all models.

   Note:

   If you are using Runtime 24.1 on Python 3.11, then you will need to change the software_spec_uid to runtime-24.1-py3.11 and the scikit-learn version to scikit-learn-1.3.

4. Run the cells in the Deploy and score section to create the online deployment, get deployment details, and send a scoring request to the deployed model to see the prediction.

5. Click File > Save File. Section 5. Create a batch deployment and score using connection asset is optional.

You can also view the model deployment directly from the deployment space. Follow these steps to test the deployed model in the space.

1. From the Navigation Menu , click Deployments.

2. Click the Spaces tab.

3. Select the appropriate deployment space from the list.

4. Click Scikit model.

5. Click Deployment of scikit model.

6. Review the Endpoint and Code snippets.

7. Click the Test tab. You can test the deployed model by pasting the following JSON code:

      {"input_data": [{"values": [[0.0, 0.0, 5.0, 16.0, 16.0, 3.0, 0.0, 0.0, 0.0, 0.0, 9.0, 16.0, 7.0, 0.0, 0.0, 0.0, 0.0, 0.0, 12.0, 15.0, 2.0, 0.0, 0.0, 0.0, 0.0, 1.0, 15.0, 16.0, 15.0, 4.0, 0.0, 0.0, 0.0, 0.0, 9.0, 13.0, 16.0, 9.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 14.0, 12.0, 0.0, 0.0, 0.0, 0.0, 5.0, 12.0, 16.0, 8.0, 0.0, 0.0, 0.0, 0.0, 3.0, 15.0, 15.0, 1.0, 0.0, 0.0], [0.0, 0.0, 6.0, 16.0, 12.0, 1.0, 0.0, 0.0, 0.0, 0.0, 5.0, 16.0, 13.0, 10.0, 0.0, 0.0, 0.0, 0.0, 0.0, 5.0, 5.0, 15.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 8.0, 15.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 13.0, 13.0, 0.0, 0.0, 0.0, 0.0, 0.0, 6.0, 16.0, 9.0, 4.0, 1.0, 0.0, 0.0, 3.0, 16.0, 16.0, 16.0, 16.0, 10.0, 0.0, 0.0, 5.0, 16.0, 11.0, 9.0, 6.0, 2.0]]}]}
   

8. Click Predict. The resulting prediction indicates that the hand-written digits are 5 and 4.