JupyterHub Service & ML Pods

Berkelium offers two primary ways to run interactive Jupyter environments:

1. Managed LBNL JupyterHub

For quick exploration and collaborative lab work, visit: 👉 jupyter.berkelium.lbl.gov

• Authenticate with your LBNL OneID.
• Select your profile:
  • Standard CPU Profile (4 cores, 16GB RAM)
  • Data Science GPU Profile (NVIDIA A100 / 32GB RAM)
  • LLM / Vision Large Profile (NVIDIA H100 80GB)

---

2. Dedicated In-Namespace JupyterLab Deployment

For labs requiring custom system packages and dedicated GPU hours, deploy a standalone JupyterLab instance inside your research namespace:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: lab-jupyter
  namespace: sci-myproject
spec:
  replicas: 1
  selector:
    matchLabels:
      app: lab-jupyter
  template:
    metadata:
      labels:
        app: lab-jupyter
    spec:
      containers:
        - name: jupyter
          image: quay.io/jupyter/pytorch-notebook:cuda12-latest
          env:
            - name: JUPYTER_TOKEN
              value: "lab-secret-token-change-me"
          ports:
            - containerPort: 8888
          resources:
            limits:
              cpu: "8"
              memory: "32Gi"
              nvidia.com/gpu: "1"
          volumeMounts:
            - name: workspace
              mountPath: /home/jovyan/work
      volumes:
        - name: workspace
          persistentVolumeClaim:
            claimName: lab-shared-data-pvc