Installation

Linux / MacOS

Conda installation:

conda install fegrow -c conda-forge

We recommend using mambaforge which is a flavour of conda. Conda struggles with resolving the environment.

To use the latest git version, download the code and use the provided environment file environment.yml

git clone https://github.com/cole-group/FEgrow.git
cd FEgrow
mamba env create -f environment.yml
conda activate fegrow
pip install --no-deps . # the repository directory

In order to ensure that the environment is available in your jupyter, you can then use

python -m ipykernel install --user --name=fegrow

Warning

Gnina stage is not supported on MacOS because CUDA's is not available on MacOS (see issue)

Windows

We have not verified whether this toolkit works on Windows and we expect issues with the dependancy openmm-torch.