Jupyter Setup on RTX 5060 Ti 16GB

JupyterLab is the default workspace for GPU-accelerated Python – notebooks, terminals, file browser and a proper IDE all in one tab. This guide walks a production-grade install on the RTX 5060 Ti 16GB via our dedicated GPU hosting – isolated env, password auth, TLS proxy and a systemd service unit that survives reboots.

Contents

• Prerequisites
• Python environment
• Password and config
• systemd service
• TLS reverse proxy
• GPU sanity check

Prerequisites

See Ubuntu driver install and first-day checklist before proceeding.

Python Environment

sudo useradd -m -s /bin/bash jupyter
sudo -iu jupyter

# Isolated venv
python3.11 -m venv ~/venv
source ~/venv/bin/activate

pip install --upgrade pip
pip install jupyterlab==4.2.* ipywidgets
pip install torch --index-url https://download.pytorch.org/whl/cu126
pip install transformers accelerate bitsandbytes peft datasets
pip install vllm  # optional, heavy

Password and Config

source ~/venv/bin/activate
jupyter lab --generate-config

# Set password (hashed)
python -c "from jupyter_server.auth import passwd; print(passwd('YOUR-PASS'))"
# argon2:$argon2id$v=19$...

cat >> ~/.jupyter/jupyter_lab_config.py << 'EOF'
c.ServerApp.ip = '127.0.0.1'              # nginx fronts it
c.ServerApp.port = 8888
c.ServerApp.open_browser = False
c.ServerApp.root_dir = '/home/jupyter/work'
c.ServerApp.password = 'argon2:...'       # paste hash here
c.ServerApp.allow_remote_access = True
c.ServerApp.token = ''                    # password only
EOF
mkdir -p ~/work

systemd Service

# /etc/systemd/system/jupyterlab.service
[Unit]
Description=JupyterLab
After=network.target

[Service]
Type=simple
User=jupyter
Group=jupyter
WorkingDirectory=/home/jupyter/work
Environment="PATH=/home/jupyter/venv/bin"
ExecStart=/home/jupyter/venv/bin/jupyter lab \
  --config=/home/jupyter/.jupyter/jupyter_lab_config.py
Restart=on-failure
RestartSec=5

[Install]
WantedBy=multi-user.target

Enable and start:

sudo systemctl daemon-reload
sudo systemctl enable --now jupyterlab
sudo systemctl status jupyterlab

TLS Reverse Proxy

# /etc/nginx/sites-available/jupyter
server {
  listen 443 ssl http2;
  server_name notebook.example.com;

  ssl_certificate     /etc/letsencrypt/live/notebook.example.com/fullchain.pem;
  ssl_certificate_key /etc/letsencrypt/live/notebook.example.com/privkey.pem;

  location / {
    proxy_pass http://127.0.0.1:8888;
    proxy_http_version 1.1;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection "upgrade";
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_read_timeout 86400;   # long-lived WS
  }
}

Get a certificate with certbot, then reload nginx. The browser loads https://notebook.example.com and prompts for the password.

GPU Sanity Check

Open a new notebook and paste:

import torch
print(torch.cuda.is_available())
print(torch.cuda.get_device_name(0))
print(f"VRAM: {torch.cuda.get_device_properties(0).total_memory / 1e9:.1f} GB")

# Tiny matmul to confirm compute
x = torch.randn(4096, 4096, device='cuda', dtype=torch.bfloat16)
y = x @ x
torch.cuda.synchronize()
print("OK")

Expected output: True, NVIDIA GeForce RTX 5060 Ti, VRAM: 16.0 GB, OK.

See also: VS Code Remote setup, Docker CUDA setup, Ubuntu driver install, first-day checklist, vLLM setup.