Notebook environments and packages
Aidin notebooks are persistent JupyterLab sessions that run on platform compute. Each notebook uses a container image and a compute profile, and keeps its files on your workspace volume.
Official image
Section titled “Official image”Every notebook currently runs Aidin’s single official notebook image:
harbor.aidin-ai.com/aidin-notebooks/kubeflow/notebooks/jupyter-scipy:f83f8313 @sha256:fcc2d7c896588e8144af66d009e416d4e844e2bbdf0c73ecca7b1aa95d198a0eInstalling packages
Section titled “Installing packages”Use standard Python tooling inside your notebook session, for example:
pip install <package>- Installed packages live in the notebook session environment. Code and data persist on your workspace volume; the package environment follows the notebook image. When a notebook is recreated or the platform moves to a new image, reinstall your packages — or keep a requirements file on your volume for a fast reset. (Pinning packages into a custom image is not currently available; the pending unified official image will address repeatable environments.)
- For Python client tooling for the Aidin platform itself, see the Python SDK.
JupyterLab
Section titled “JupyterLab”JupyterLab behaves as standard JupyterLab: notebooks, terminals, and file browser, with your workspace volume mounted at /home/jovyan. You can also use TensorBoard alongside notebooks — see Integrations.
Session behavior (verified 2026-11-10):
- Stop/start preserves all files on your workspace volume.
- Sessions stop automatically after 30 minutes of inactivity or 24 hours of runtime; restart from the dashboard at any time.
- Compute is released while a notebook is stopped.