Skip to content

Installation & Setup

Jump to bottom
Durand edited this page Dec 1, 2025 · 1 revision

Installation & Setup: Guide to Prerequisites and Installation

This comprehensive guide details the necessary system prerequisites and provides step-by-step instructions for installing the DeepDock API on your host system.

⚠️ Important Note on Operating System

DeepDock is currently developed and tested primarily for Linux environments. While some components may function on other systems, full support, including for Windows, is planned for future releases. It is highly recommended to run DeepDock on a compatible Linux distribution.

1. Prerequisites Check and Detailed Installation

DeepDock is designed to manage GPU-aware containers, meaning your host system must have the correct ecosystem installed before starting the DeepDock service.

1.1 Docker Engine

Docker is the foundation for container management. DeepDock relies entirely on the Docker daemon for creating, managing, and inspecting containers.

  • Requirement: Docker Engine must be installed and running.
  • Installation Guide: Docker Installer
  • Verification: Run the command below to ensure Docker is accessible:
docker --version

If installation is successful, this should return the Docker version number.

1.2 NVIDIA GPU and Drivers

DeepDock's core functionality is tied to the presence of an NVIDIA GPU and its corresponding drivers, which are necessary for the NVML monitoring feature.

  • Requirement: A compatible NVIDIA GPU and the latest proprietary driver package installed.
  • Driver Download: Nvidia Drivers homepage
  • Verification: Run the NVIDIA System Management Interface (SMI) utility:
nvidia-smi

A successful output should display your GPU status, driver version, and current utilization (e.g., temperature, memory usage).

1.3 NVIDIA Container Toolkit

This toolkit is critical. It allows Docker to communicate with the NVIDIA drivers and attach the necessary GPU resources to your running containers, enabling GPU access inside the container.

  • Requirement: The NVIDIA Container Toolkit must be installed and configured for Docker integration.
  • Installation Guide: NVIDIA Container Toolkit Docs
  • Verification: Test the low-level CLI tool provided by the toolkit:
nvidia-container-cli --version

This confirms the toolkit itself is installed. To verify Docker integration, you can run a GPU test container:

docker run --rm --gpus all nvidia/cuda:12.1.0-base nvidia-smi

(This command should succeed and show the GPU output from inside the container.)

2. DeepDock Installation

Once all system prerequisites are verified, DeepDock can be installed as a standard Python package via pip.

2.1 Python and Pip Requirement

DeepDock requires Python 3.8 or higher. The package manager pip is essential for installation.

  • Requirement: Python 3.8+ and pip must be installed.
  • Installation Note: If Python or pip are not installed, please follow the official Python documentation for your specific operating system (e.g., using apt, yum, or brew).
  • Verification:
python3 --version
pip --version

2.2 Package Installation

Install DeepDock directly from its GitHub repository:

pip install git+[https://github.com/JuniorDurand/DeepDock.git](https://github.com/JuniorDurand/DeepDock.git)

This command installs the DeepDock API (FastAPI) and all associated dependencies needed for container and NVML interaction.

3. Post-Installation: Running the Service

After installation, you can launch the DeepDock API service using the command-line entry point.

deepdock

Clone this wiki locally