To spin up a notebook in your platform, follow these steps:
- Go to the Notebooks tab in the deployment dashboard.
- Click on "New Deployment" to start creating a new notebook.
- Select the desired workspace where you want to create the notebook.
- Click on the "Next Step" button.
Once you've completed these initial steps, you can proceed to fill in the settings for your notebook using the provided parameters.
Parameters for Notebook Configuration:
- Name: This parameter uniquely identifies the notebook in the workspace.
- Image Type: Choose a notebook image to deploy. There are two options available:
- TruefoundryImageBase: JupyterLab with a persistent Python environment (Python 3.8.10).
- TruefoundryImageFull: JupyterLab with all common ML libraries and a persistent environment (Python 3.8.10).
- Enable Root Access (optional): Enable root access to the container. Note that changes made to the root directory / will not be persisted across notebook restarts.
- List of Debian packages to install (optional): Specify a list of Debian packages to install via apt get. Each package should be a string.
- Endpoint: Specify the endpoint for the notebook. This determines the URL or address where the notebook will be accessible.
- Host: Specify the host for the notebook endpoint. This can be a domain name or IP address, such as "ai.example.com" or "app.truefoundry.com".
- Path (optional): You can specify a path for the notebook endpoint, such as "/v1/api/ml/" or "/v2/docs/".
- Auth: Add Basic Auth credentials to protect the notebook endpoint if desired.
- Cull Timeout: Set the duration of inactivity (in minutes) after which the notebook instance will be stopped. Even if the notebook is open in the browser, if no activity is detected, it will be stopped.
- Resources: Configure the resource constraints for the notebook. This includes parameters such as CPU and memory requests and limits, ephemeral storage requests and limits, GPU count (if applicable), and shared memory size.
- Home Directory Size: Specify the size of the home directory for the notebook in GB. This represents persistent storage for the notebook.
- Service Account (optional): You can specify a service account that the notebook workload should use.
- Mounts: Configure data to be mounted to the notebook pod(s). This parameter allows you to mount secrets, string data, or volumes to the notebook environment.
After configuring the desired parameters, proceed with clicking the Submit button, the deployment process will now begin.
Once the deployment is successful and the status of the notebook changes to "Active" with the status "Running," you will be able to click on the endpoint of the notebook.
There might be a slight delay before the notebook becomes live and accessible after the status shows "Running." Please be patient and wait for a few seconds before attempting to access the notebook.
Voila!! Now you can access the notebook
Once you have finished working with the notebook, you can stop it to save resources and reduce costs.
When you click the stop button, the system will initiate the process of culling the notebook.
This process involves shutting down the notebook environment and releasing the associated resources. By stopping the notebook, you effectively suspend its operation, allowing you to conserve resources and avoid unnecessary costs when the notebook is not in use.
Your data (apart from the apt packages installed as a root user) will be persisted, and you can easily restart it at a later time by clicking the resume button.
Updated 25 days ago