:py:mod:`tlm_adjoint.torch`
===========================

.. py:module:: tlm_adjoint.torch

.. autoapi-nested-parse::

   Interface with PyTorch.

   Can be used to embed models, differentiated by tlm_adjoint, within a PyTorch
   calculation. Follows the same principles as described in

       - Nacime Bouziani and David A. Ham, 'Physics-driven machine learning models
         coupling PyTorch and Firedrake', 2023, arXiv:2303.06871v3

   ..
       !! processed by numpydoc !!


Module Contents
---------------

.. py:function:: to_torch_tensors(X, *args, **kwargs)

   
   Convert one or more variables to :class:`torch.Tensor` objects.


   :Parameters:

       **X** : variable or Sequence[variable, ...]
           Variables to be converted.

       **args, kwargs**
           Passed to :func:`torch.tensor`.



   :Returns:

       tuple[variable, ...]
           The converted variables.











   ..
       !! processed by numpydoc !!

.. py:function:: from_torch_tensors(X, X_t)

   
   Copy data from PyTorch tensors into variables.


   :Parameters:

       **X** : variable or Sequence[variable, ...]
           Output.

       **X_t** : Sequence[:class:`torch.Tensor`, ...]
           Input.














   ..
       !! processed by numpydoc !!

.. py:function:: torch_wrapped(forward, space, *, manager=None, clear_caches=True)

   
   Wrap a model, differentiated using tlm_adjoint, so that it can be used
   with PyTorch.


   :Parameters:

       **forward** : callable
           Accepts one or more variable arguments, and returns a variable or
           :class:`Sequence` of variables.

       **space** : space or Sequence[space, ...]
           Defines the spaces for input arguments.

       **manager** : :class:`.EquationManager`
           Used to create an internal manager via :meth:`.EquationManager.new`.
           `manager()` is used if not supplied.

       **clear_caches** : bool
           Whether to clear caches before a call of `forward`.



   :Returns:

       callable
           A version of `forward` with :class:`torch.Tensor` inputs and outputs.











   ..
       !! processed by numpydoc !!