Skip to content

Latest commit

 

History

History
25 lines (20 loc) · 1.32 KB

conda-instructions.md

File metadata and controls

25 lines (20 loc) · 1.32 KB
Raw

Intstructions to Install Conda Packages

In this file, I'll list out all of the packages that you need to install via conda. For each package, I'll give it in the form "X:x", where X is the package name and x is what you type in the conda command to install, i.e. conda install x.

  1. PyTorch: torch (installation instructions)
  • Note, make sure you install for your correct CPU/GPU version. This only really matters for matching cuda versions on the GPU, CPU install instructions should be straightforward.
  1. PyTorch Geometric and corresponding packages: pyg (installation instructions)
  • Using conda install pyg should take care of version conflicts, so long as you first install torch
  1. Rdkit: rdkit (installation instructions, see here also)

Additional pip requirements (latest version of each should work):

  1. pandas
  2. numpy
  3. tqdm (progress bar)
  4. matplotlib
  5. sklearn
  6. scipy
  7. captum (tools for explainability)
  8. networkx

Finally, you'll need to make sure that the reference to the polymerlearn package is installed locally. From the main PolymerGNN directory, please run

pip install -e .