Installation

DeepReg is written in Python 3.7. Dependent external libraries provide core IO functionalities and other standard processing tools. The dependencies for DeepReg are managed by pip.

The package is primarily distributed via PyPI.

Create a virtual environment

The recommended method is to install DeepReg in a dedicated virtual environment to avoid issues with other dependencies. The conda environment is recommended:

Anaconda / Miniconda

DeepReg primarily supports and is regularly tested with Ubuntu and Debian Linux distributions.

Linux

Mac OS

Windows

With CPU only

conda create --name deepreg python=3.7 tensorflow=2.2
conda activate deepreg

With GPU

conda create --name deepreg python=3.7 tensorflow-gpu=2.2
conda activate deepreg

With CPU only

conda create --name deepreg python=3.7 tensorflow=2.2
conda activate deepreg

With GPU

Warning

Not supported or tested.

With CPU only

Warning

DeepReg on Windows is not fully supported. However, you can use the Windows Subsystem for Linux with CPU only. Set up WSL and follow the DeepReg setup instructions for Linux.

With GPU

Warning

Not supported or tested.

Install the package

Install from the cloned local project

git clone https://github.com/DeepRegNet/DeepReg.git # clone the repository
cd DeepReg # change working directory to the DeepReg root directory
pip install -e . # install the package

Install from the PyPI release

pip install deepreg

Note

All dependencies, APIs and command-line tools will be installed automatically via either installation method. However, the PyPI release currently does not ship with test data and demos. Running examples in this documentation may require downloading test data and changing default paths to user-installed packages with the PyPI release. These examples include those in the Quick Start and DeepReg Demo