In this hands-on lab, you will become familiar with performing the basic actions for hyperparameter tuning.
Successfully complete this lab by achieving the following learning objectives:
- Create a Compute Resource and Access the Provided Jupyter Notebook
Utilize the provisioned workspace to create a compute instance. This will be used to power a Jupyter notebook and execute your machine learning tasks.
If you plan to utilize the existing Jupyter notebook with the data and steps required to create and run the experiment already preconfigured, clone the repo from GitHub.
Use the Standard_DS2_v2 VM size.
- Simulate Sample Data
Generate random data to simulate the diabetes patient information.
- Create a Training Script Using Hyperparameters
Prepare a training script that will train the model using hyperparameters.
- Find and Register the Best Model
Run a hyperdrive experiment to find the best performing model, and then register it.