Custom training autologging: Notebook

As a data scientist experimenting with large models, you need a way to run experiments on a scalable training service to log parameters and metrics. This enables reproducibility.

With Vertex AI training and experiments autologging integration, you can run your ML experiments at scale and autolog their parameters and metrics by using the enable_autolog argument.

Notebook: Vertex AI Experiments: Custom training autologging - Local script

To see an example of getting started with custom autologging with a local script, run the "Vertex AI Experiments: Autologging" notebook in one of the following environments:

Open in Colab | Open in Colab Enterprise | Open in Vertex AI Workbench | View on GitHub

This tutorial uses the following Google Cloud ML services and resources:

• Vertex AI Experiments
• Vertex AI training

The steps performed include:

1. Formalize model experiment in a script.
2. Run model training using local script on Vertex AI training.
3. Check out ML experiment parameters and metrics in Vertex AI Experiments.

Relevant content

• Run training job with experiment tracking