Vertex AI Workbench is a JupyterLab notebook-based development environment available for your entire data science workflow. You can interact with Vertex AI and its services on Google Distributed Cloud (GDC) air-gapped from within a notebook of a JupyterLab instance that Vertex AI Workbench provides.
Vertex AI Workbench integrations and features make accessing your machine learning data easier, sharing and processing data faster, interacting with Vertex AI services using the Python programming language, and more.
For example, Vertex AI Workbench lets you do the following:
- Access and explore your machine learning data from within a JupyterLab notebook.
- Share your JupyterLab notebook with other users of your project.
- Import Vertex AI client libraries to simplify accessing APIs programmatically.
- Interact with Vertex AI services, authenticate API requests, and use Vertex AI features from Python scripts.
- Create a backup and restore your JupyterLab instance data.
- Use an integrated development environment (IDE) to use built-in integrations of JupyterLab notebooks.
- Set up an end-to-end notebook-based production environment.
JupyterLab instances
Vertex AI Workbench offers JupyterLab instances with built-in integrations that help you set up an end-to-end notebook-based production environment. JupyterLab instances combine workflow-oriented integrations of a managed instance with the customization and control you need over your environment.
Vertex AI Workbench includes instance types preinstalled with JupyterLab and a suite of deep learning packages, including support for the TensorFlow and PyTorch frameworks. Depending on your needs, you can choose between CPU-only or GPU-enabled instances.
You can select a Docker image and a cluster for your JupyterLab instance environment. Docker lets you create a custom JupyterLab environment and build it into an image. This image ensures consistency and reproducibility across different deployments, including all the necessary packages and tools. You can share this customized environment with others or use it as a foundation for future development.
JupyterLab instances are protected by authentication and authorization.