.. _architecture-graph2mat:

Graph2Mat
=========

Interface of ``Graph2Mat`` to all architectures in ``metatrain``.

.. _architecture-graph2mat_installation:

Installation
------------

To install this architecture along with the ``metatrain`` package, run:

.. code-block:: bash

    pip install metatrain[graph2mat]

where the square brackets indicate that you want to install the optional
dependencies required for ``graph2mat``.

.. _architecture-graph2mat_default_hypers:

Default Hyperparameters
-----------------------

The description of all the hyperparameters used in ``graph2mat`` is provided
further down this page. However, here we provide you with a yaml file containing all
the default hyperparameters, which might be convenient as a starting point to
create your own hyperparameter files:

.. literalinclude:: ../default_hypers/graph2mat-default-hypers.yaml
   :language: yaml

.. _architecture-graph2mat_model_hypers:

Model hyperparameters
------------------------

The parameters that go under the ``architecture.model`` section of the config file
are the following:

.. container:: mtt-hypers-remove-classname

    ..

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.ModelHypers.basis_yaml

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.ModelHypers.basis_grouping

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.ModelHypers.node_hidden_irreps

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.ModelHypers.edge_hidden_irreps



.. _architecture-graph2mat_trainer_hypers:

Trainer hyperparameters
-------------------------

The parameters that go under the ``architecture.trainer`` section of the config file
are the following:

.. container:: mtt-hypers-remove-classname

    ..

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.optimizer

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.optimizer_kwargs

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.lr_scheduler

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.lr_scheduler_kwargs

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.distributed

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.distributed_port

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.batch_size

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.num_epochs

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.log_interval

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.checkpoint_interval

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.per_structure_targets

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.num_workers

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.log_mae

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.log_separate_blocks

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.best_model_metric

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.grad_clip_norm

        .. autoattribute:: metatrain.experimental.graph2mat.documentation.TrainerHypers.loss



.. _architecture-graph2mat_references:

References
----------

.. footbibliography::