Skip to content

Latest commit

 

History

History
60 lines (37 loc) · 1.46 KB

README.rst

File metadata and controls

60 lines (37 loc) · 1.46 KB

Overview

This is the home of the cormorant software package for learning on atomic environments.

Documentation

To install the documentation, go to the docs folder and run "make html". You will need to install the sphinx_rtd_theme (this can be done using pip install).

Getting started

Installation

Cloning the git repo

Cormorant can be cloned directly from the git repo using:

git clone https://github.com/risilab/cormorant.git
Using pip

Cormorant is installable from source using pip. You can currently install it from source by going to the directory with setup.py and running:

pip install cormorant .

If you would like to modify the source code directly, note that Cormorant can also be installed in "development mode" using the command:

pip install cormorant -e .

Training example

The example training script is in :examples/train_cormorant.py:. The same script can train both the datasets QM9 and MD17, and can also be extended to more general datasets.

python examples/train_qm9.py
python examples/train_md17.py

Note that if no GPU is available, the the training script will throw an error. To force CPU-based training, add the : --cpu: flag

Architecture

A more detailed description of the Cormorant architecture is available in the Cormorant paper.