- Navigate to Google Cloud Console using this URL: https://console.cloud.google.com/welcome
- In the search bar at the top of the console, type "Vertex AI" and select it from the search results
- Within Vertex AI, locate and select "Workbench" from the available options
- Click "Create New Instance" to begin setting up your Jupyter environment
- Basic Setup:
- Provide a meaningful name for your instance
- Click "Advanced Options" to access additional configuration settings
- Environment Configuration:
- Navigate to the "Environment" tab
- Select "JupyterLab 4" as your development environment
-
Look for the "Startup Script" option
-
Click "Browse" to locate the script
-
Paste to "gs://nigms-sandbox/nosi-duke/post_start_clean_032225.sh"
- Machine Configuration:
- Click "Create" at the bottom of the page
- Wait approximately 3-5 minutes for the instance to be fully provisioned
- Once provisioning is complete, click "Start JupyterLab" to launch the environment
- In JupyterLab, locate and click the Git button in the sidebar
-
Select "Clone a Repository"
-
Enter the following repository URL: https://github.com/mitomac/duke_nigms_sandbox.git
-
Click "Clone" to download the repository
- Navigate into the project by double-clicking the "duke_nigms_sandbox" folder to access the modules
-
If you're working with R notebooks, you'll be prompted to select a kernel
-
Choose the R kernel when prompted to ensure proper execution of R code
-
Start the AI assistant
-
Generate and copy API key by running through Module 0 notebook.
-
Select a google gemini model and input the API key you just generated.
Additional Tips:
-
Keep track of your instance's running time to manage resources effectively
-
Save your work regularly
-
Consider bookmarking the Google Cloud Console URL for easier access in the future
