Hyperparameter Optimization using Notebooks

Tuning your model to get the best results? Here’s how to set up hyperparameter optimization using the Notebook:

• Define Parameters and Inputs: Make sure you set up your model’s parameters and inputs directly in your notebook.

Example of defining inputs and parameters in your Notebook:

import valohai
valohai.prepare(
    step='<your-step-name>',
    image='<your-docker-image:tag>',
    default_inputs={
        'input-name': '<input-url>'
    },
    default_parameters={
        'param1': value1,
        'param2': value2
    }
)

• Run and Version Your Job: Use the “Run Remote” button to execute and version your entire notebook. This will create a separate execution from your Notebook.
• Convert to Task: After a successful run, turn your execution into a task for further analysis and optimization by navigating to the completed execution in the Valohai UI and clicking the Task icon.

When creating a task, you can select the Task Type and configure the tuning parameters. Valohai will then create multiple executions to test the specified parameters for you.

With these steps, you’ll be able to make the most of Valohai Hosted Notebooks, simplifying and enhancing your machine learning workflow!