Environment setup

Git clone the repo

To get started with the code, please create a new conda environment as follows :

  conda env create -f env_gpu.yml 
  conda activate graphino  # graphino is the name of the environment

This will install all required dependencies, uses Python 3.7, and expects CUDA to be available. If no GPU is available, please remove line 7 (cudatoolkit) from the environment file, and proceed as above.
This should be enough for Pycharm, for command line stuff you may need to then also run python setup.py install.

Alternatives/If you encounter an error

If you encounter, e.g., a cryptic InvalidArchiveError, please try instead with

conda create -n graphino python=3.7 conda activate graphino conda install pytorch torchvision torchaudio cudatoolkit=10.2 -c pytorch # *** pip install -r requirements.txt

*** look in https://pytorch.org/ for the exact command (which depends on your CUDA version).

If this does not work either, do not hesitate in contacting us (although it should work).