.. DO NOT EDIT.
.. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY.
.. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE:
.. "_auto_examples/kernel_ridge/plot_model_on_gpu.py.py"
.. LINE NUMBERS ARE GIVEN BELOW.

.. only:: html

    .. note::
        :class: sphx-glr-download-link-note

        Click :ref:`here <sphx_glr_download__auto_examples_kernel_ridge_plot_model_on_gpu.py.py>`
        to download the full example code

.. rst-class:: sphx-glr-example-title

.. _sphx_glr__auto_examples_kernel_ridge_plot_model_on_gpu.py.py:


Fitting a model on GPU
======================

This example demonstrates how to fit a model using GPU computations.

Himalaya implements different computational backends to fit the models:

- "numpy" (CPU) (default)
- "torch" (CPU)
- "torch_cuda" (GPU)
- "cupy" (GPU)

Each backend is only available if you installed the corresponding package with
CUDA enabled. Check the ``pytorch``/``cupy`` documentation for installation
instructions.

.. GENERATED FROM PYTHON SOURCE LINES 20-22

Create a random dataset
-----------------------

.. GENERATED FROM PYTHON SOURCE LINES 22-27

.. code-block:: default

    import numpy as np
    n_samples, n_features, n_targets = 10, 20, 4
    X = np.random.randn(n_samples, n_features)
    Y = np.random.randn(n_samples, n_targets)








.. GENERATED FROM PYTHON SOURCE LINES 28-34

Change backend
--------------
To change the backend, you need to call the function
``himalaya.backend.set_backend``. With the option ``on_error="warn"``, the
function does not raise an error if the new backend fails to be imported, and
the backend is kept unchanged.

.. GENERATED FROM PYTHON SOURCE LINES 34-38

.. code-block:: default


    from himalaya.backend import set_backend
    backend = set_backend("cupy", on_error="warn")








.. GENERATED FROM PYTHON SOURCE LINES 39-47

GPU backend
-----------
To fit a himalaya model on GPU, you don't need to move the input arrays to
GPU, the method ``fit`` will do it for you. However, the float precision will
not be changed.

To make the most of GPU memory and computational speed, you might want to
change the float precision to float32.

.. GENERATED FROM PYTHON SOURCE LINES 47-53

.. code-block:: default

    X = X.astype("float32")

    from himalaya.kernel_ridge import KernelRidge
    model_him = KernelRidge(kernel="linear", alpha=0.1)
    model_him.fit(X, Y)





.. rst-class:: sphx-glr-script-out

 Out:

 .. code-block:: none


    KernelRidge(alpha=0.1)



.. GENERATED FROM PYTHON SOURCE LINES 54-57

The results are stored in GPU memory, using an array object specific to the
backend used. To use the results in other libraries (for example matplotlib),
you can create a numpy array using the function ``backend.to_numpy``.

.. GENERATED FROM PYTHON SOURCE LINES 57-61

.. code-block:: default

    scores = model_him.score(X, Y)
    print(scores.__class__)
    scores = backend.to_numpy(scores)
    print(scores.__class__)




.. rst-class:: sphx-glr-script-out

 Out:

 .. code-block:: none

    <class 'cupy.core.core.ndarray'>
    <class 'numpy.ndarray'>





.. rst-class:: sphx-glr-timing

   **Total running time of the script:** ( 0 minutes  0.446 seconds)


.. _sphx_glr_download__auto_examples_kernel_ridge_plot_model_on_gpu.py.py:


.. only :: html

 .. container:: sphx-glr-footer
    :class: sphx-glr-footer-example



  .. container:: sphx-glr-download sphx-glr-download-python

     :download:`Download Python source code: plot_model_on_gpu.py.py <plot_model_on_gpu.py.py>`



  .. container:: sphx-glr-download sphx-glr-download-jupyter

     :download:`Download Jupyter notebook: plot_model_on_gpu.py.ipynb <plot_model_on_gpu.py.ipynb>`


.. only:: html

 .. rst-class:: sphx-glr-signature

    `Gallery generated by Sphinx-Gallery <https://sphinx-gallery.github.io>`_