Workspaces

A Workspace is an interactive development environment with GPU access. It runs a Jupyter notebook server that you can access from your browser.

Creating a Workspace

1. Go to Workspaces in the Run:AI UI
2. Click New Workspace
3. Configure:
4. Name: Give it a descriptive name (e.g., my-jupyter)
5. Project: Select your assigned project
6. Environment: Choose from the available environments
7. Compute Resource: Select GPU type and count

8. Data Sources: Attach any datasets you need

9. Click Create

Accessing Your Workspace

Once the workspace is running:

1. Click on the workspace name in the list
2. Click the Connect button
3. Select Jupyter to open the notebook interface

Workspace Lifecycle

Status	Meaning
Pending	Waiting for GPU allocation
Running	Active and accessible
Stopped	Manually stopped (data preserved)
Failed	Error occurred during startup

Tips

• Save your work frequently - Workspaces can be preempted if you exceed your GPU quota
• Stop workspaces when not in use - This frees GPUs for other students
• Use /home/user/ for persistent storage - Files outside this path are lost on restart

If your workspace stays in "Pending" for more than 5 minutes, ask an instructor to check GPU availability.