Graph Neural Networks
======================

Below, the implemented message passing algorithms and message passing layers,
as well as a few complete models are described. All of these rely on `PyTorch <https://pytorch.org/docs/stable/index.
html>`_ and `PyTorch Geometric <https://pytorch-geometric.readthedocs.io/en/latest/>`_
as backbones.

.. contents:: Contents
    :local:

Weave
-------

.. automodule:: grape_chem.models.Weave_gnn
   :members:
   :undoc-members:
   :show-inheritance:

MPNN
-------

.. automodule:: grape_chem.models.MPNN_gnn
   :members:
   :undoc-members:
   :show-inheritance:

D-MPNN
-------

.. automodule:: grape_chem.models.DMPNN_gnn
   :members:
   :undoc-members:
   :show-inheritance:


MEGNet
-------

.. automodule:: grape_chem.models.MEGNet_gnn
   :members:
   :undoc-members:
   :show-inheritance:

AFP
-----

.. automodule:: grape_chem.models.AFP_gnn
   :members:
   :undoc-members:
   :show-inheritance:



Output Model
---------------

.. automodule:: grape_chem.models.test_models
   :members:
   :undoc-members:
   :show-inheritance: