.. DO NOT EDIT.
.. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY.
.. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE:
.. "gallery_2d/mnist_keras.py"
.. LINE NUMBERS ARE GIVEN BELOW.

.. only:: html

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

        Click :ref:`here <sphx_glr_download_gallery_2d_mnist_keras.py>`
        to download the full example code

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

.. _sphx_glr_gallery_2d_mnist_keras.py:


Classification of MNIST with scattering
=======================================
Here we demonstrate a simple application of scattering on the MNIST dataset.
We use 10000 images to train a linear classifier. Features are normalized by
batch normalization.

.. GENERATED FROM PYTHON SOURCE LINES 10-15

Preliminaries
-------------

Since we're using TensorFlow and Keras to train the model, import the
relevant modules.

.. GENERATED FROM PYTHON SOURCE LINES 15-20

.. code-block:: default


    import tensorflow as tf
    from tensorflow.keras.models import Model
    from tensorflow.keras.layers import Input, Flatten, Dense


.. GENERATED FROM PYTHON SOURCE LINES 21-23

Finally, we import the `Scattering2D` class from the `kymatio.keras`
package.

.. GENERATED FROM PYTHON SOURCE LINES 23-26

.. code-block:: default


    from kymatio.keras import Scattering2D


.. GENERATED FROM PYTHON SOURCE LINES 27-31

Training and testing the model
------------------------------

First, we load in the data and normalize it.

.. GENERATED FROM PYTHON SOURCE LINES 31-36

.. code-block:: default


    (x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()

    x_train, x_test = x_train / 255., x_test / 255.


.. GENERATED FROM PYTHON SOURCE LINES 37-39

We then create a Keras model using the scattering transform followed by a
dense layer and a softmax activation.

.. GENERATED FROM PYTHON SOURCE LINES 39-46

.. code-block:: default


    inputs = Input(shape=(28, 28))
    x = Scattering2D(J=3, L=8)(inputs)
    x = Flatten()(x)
    x_out = Dense(10, activation='softmax')(x)
    model = Model(inputs, x_out)


.. GENERATED FROM PYTHON SOURCE LINES 47-48

Display the created model.

.. GENERATED FROM PYTHON SOURCE LINES 48-51

.. code-block:: default


    model.summary()


.. GENERATED FROM PYTHON SOURCE LINES 52-54

Once the model is created, we couple it with an Adam optimizer and a
cross-entropy loss function.

.. GENERATED FROM PYTHON SOURCE LINES 54-59

.. code-block:: default


    model.compile(optimizer='adam',
                  loss='sparse_categorical_crossentropy',
                  metrics=['accuracy'])


.. GENERATED FROM PYTHON SOURCE LINES 60-61

We then train the model using `model.fit` on a subset of the MNIST data.

.. GENERATED FROM PYTHON SOURCE LINES 61-65

.. code-block:: default


    model.fit(x_train[:10000], y_train[:10000], epochs=15,
              batch_size=64, validation_split=0.2)


.. GENERATED FROM PYTHON SOURCE LINES 66-67

Finally, we evaluate the model on the held-out test data.

.. GENERATED FROM PYTHON SOURCE LINES 67-69

.. code-block:: default


    model.evaluate(x_test, y_test)


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

   **Total running time of the script:** ( 0 minutes  0.000 seconds)


.. _sphx_glr_download_gallery_2d_mnist_keras.py:

.. only:: html

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


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

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

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

      :download:`Download Jupyter notebook: mnist_keras.ipynb <mnist_keras.ipynb>`


.. only:: html

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

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