GPU-Enabled JupyterLab Container

Docker setup for JupyterLab with full NVIDIA GPU acceleration support. Ideal for machine learning, deep learning, and scientific computing workflows. Pre-configured with CUDA libraries and Python data science stack.

Add screenshot link here

✨ Features

• CUDA 12.2 base image for full GPU acceleration
• Persistent storage for notebooks and datasets
• Secure authentication with password protection
• Resource isolation for GPU devices
• Pre-installed common Python ML packages
• Customizable environment via Dockerfile
• Docker Compose for single-command management

⚙️ System Requirements

• Docker Engine 20.10+
• NVIDIA Container Toolkit (Install Guide)
• NVIDIA Drivers 525.60+ (Driver Download)
• Docker Compose 2.0+
• Minimum 10GB free disk space

🚀 Setup Guide

1. Configure Environment

Create password file (replace your_secure_password):

echo "JUPYTER_PASSWORD=your_secure_password" > .env

2. Build Docker Image (Mandatory Step)

docker compose build --no-cache

Note: --no-cache ensures fresh dependencies install. Remove for faster rebuilds when not changing packages.

3. Launch Container

docker compose up -d

4. Access JupyterLab

Open in browser:

http://localhost:8888

🔑 Password: your_secure_password (from .env file)

⚙️ Configuration

File	Purpose	Customization
docker-compose.yml	Container definition	Port mapping, volume locations, GPU limits
requirements.txt	Python packages	Add/remove Python dependencies
Dockerfile	Image build recipe	System packages, environment variables
.env	Secrets	Jupyter password

Common Customizations:

Change Jupyter port (edit docker-compose.yml):

ports:
  - "8999:8888"  # Host:Container port

Add Python packages (edit requirements.txt):

torch==2.1.0
transformers==4.35
xgboost==2.0.0

Limit GPU access (edit docker-compose.yml):

deploy:
  resources:
    reservations:
      devices:
        - driver: nvidia
          count: 1  # Use only 1 GPU
          capabilities: [gpu]

🛠️ Management Commands

Command	Action
docker compose logs -f	View real-time logs
docker compose exec jupyter bash	Enter container shell
docker compose restart	Restart service
docker compose down	Stop container
docker compose pull	Update base images

⚠️ Troubleshooting

GPU not detected? Verify:

• nvidia-smi shows available GPUs on host
• NVIDIA Container Toolkit is properly installed
• Docker has GPU access:

docker run --rm --gpus all nvidia/cuda:12.2.0-base nvidia-smi

Port conflict? Change host port mapping in docker-compose.yml:

ports:
  - "NEW_PORT:8888"  # e.g., 9999:8888

📂 Directory Structure

.
├── Dockerfile             # Image build instructions
├── docker-compose.yml     # Container configuration
├── requirements.txt       # Python dependencies
├── workspace/             # Persistent notebooks (auto-created)
└── .env                   # Environment variables