svendsenorway/arctic-stack repository overview

⁠Arctic Stack: Advanced AI & ML Docker Environment

Arctic Stack is a Dockerized solution crafted for cutting-edge AI and machine learning development running on Ubuntu 22.04. It integrates NVIDIA's CUDA 12.3.1 toolkit providing a sophisticated platform for GPU-accelerated computational tasks.

⁠Technical Specifications:

• Base Image: nvidia/cuda:12.3.1-devel-ubuntu22.04.
• GPU Compatibility: Configured for NVIDIA GPUs, leveraging CUDA for high-performance machine learning applications. Includes both CUDA & cuDNN.

⁠Core Libraries:

• TensorFlow: Robust framework for deep learning applications.
• PyTorch: Flexible tool for AI research and production.
• Keras: High-level neural networks API, running on top of TensorFlow.
• JAX: For high-performance machine learning research.

The image also contains relevant Python libraries, such as Numpy, SciPy, Pandas for numerical and data processing tasks as well as Matplotlib & Seaborn for visualization.

⁠Additional Features:

• Jupyter Notebook: Integrated for interactive development, data analysis, and visualization.
• NVIDIA Docker Integration: Environment variables (NVIDIA_VISIBLE_DEVICES, NVIDIA_DRIVER_CAPABILITIES, NVIDIA_REQUIRE_CUDA) set for optimal utilization of NVIDIA Docker runtime.

⁠Usage:

Optimized for systems with NVIDIA GPUs. Ensure NVIDIA drivers and NVIDIA Container Toolkit are installed on the host system for full functionality.

Getting Started:

# Pull the Docker image
docker pull yourdockerhubusername/arctic-stack:latest

# Run the container with NVIDIA GPU support
docker run --gpus all -it yourdockerhubusername/arctic-stack bash

Designed for data scientists, AI researchers, and software developers, the Arctic Stack provides a sophisticated, ready-to-use environment for diverse machine learning and deep learning projects.