Python tensorflow.keras.backend.backend() Examples

def get_dropout(**kwargs):
    """Wrapper over custom dropout. Fix problem of ``None`` shape for tf.keras.
    It is not possible to define FixedDropout class as global object,
    because we do not have modules for inheritance at first time.

    Issue:
        https://github.com/tensorflow/tensorflow/issues/30946
    """
    class FixedDropout(layers.Dropout):
        def _get_noise_shape(self, inputs):
            if self.noise_shape is None:
                return self.noise_shape

            symbolic_shape = backend.shape(inputs)
            noise_shape = [symbolic_shape[axis] if shape is None else shape
                           for axis, shape in enumerate(self.noise_shape)]
            return tuple(noise_shape)

    return FixedDropout 

def swish(x):
    """Swish activation function.
    # Arguments
        x: Input tensor.
    # Returns
        The Swish activation: `x * sigmoid(x)`.
    # References
        [Searching for Activation Functions](https://arxiv.org/abs/1710.05941)
    """
    if K.backend() == 'tensorflow':
        try:
            # The native TF implementation has a more
            # memory-efficient gradient implementation
            return K.tf.nn.swish(x)
        except AttributeError:
            pass

    return x * K.sigmoid(x) 

def correct_pad(backend, inputs, kernel_size):
    """Returns a tuple for zero-padding for 2D convolution with downsampling.
    # Arguments
        input_size: An integer or tuple/list of 2 integers.
        kernel_size: An integer or tuple/list of 2 integers.
    # Returns
        A tuple.
    """
    img_dim = 2 if backend.image_data_format() == 'channels_first' else 1
    input_size = backend.int_shape(inputs)[img_dim:(img_dim + 2)]

    if isinstance(kernel_size, int):
        kernel_size = (kernel_size, kernel_size)

    if input_size[0] is None:
        adjust = (1, 1)
    else:
        adjust = (1 - input_size[0] % 2, 1 - input_size[1] % 2)

    correct = (kernel_size[0] // 2, kernel_size[1] // 2)

    return ((correct[0] - adjust[0], correct[0]),
            (correct[1] - adjust[1], correct[1])) 

def _se_block(inputs, filters, se_ratio, prefix):
    x = GlobalAveragePooling2D(name=prefix + 'squeeze_excite/AvgPool')(inputs)
    if K.image_data_format() == 'channels_first':
        x = Reshape((filters, 1, 1))(x)
    else:
        x = Reshape((1, 1, filters))(x)
    x = Conv2D(_depth(filters * se_ratio),
                      kernel_size=1,
                      padding='same',
                      name=prefix + 'squeeze_excite/Conv')(x)
    x = ReLU(name=prefix + 'squeeze_excite/Relu')(x)
    x = Conv2D(filters,
                      kernel_size=1,
                      padding='same',
                      name=prefix + 'squeeze_excite/Conv_1')(x)
    x = Activation(hard_sigmoid)(x)
    #if K.backend() == 'theano':
        ## For the Theano backend, we have to explicitly make
        ## the excitation weights broadcastable.
        #x = Lambda(
            #lambda br: K.pattern_broadcast(br, [True, True, True, False]),
            #output_shape=lambda input_shape: input_shape,
            #name=prefix + 'squeeze_excite/broadcast')(x)
    x = Multiply(name=prefix + 'squeeze_excite/Mul')([inputs, x])
    return x 

def correct_pad(backend, inputs, kernel_size):
    """Returns a tuple for zero-padding for 2D convolution with downsampling.
    # Arguments
        input_size: An integer or tuple/list of 2 integers.
        kernel_size: An integer or tuple/list of 2 integers.
    # Returns
        A tuple.
    """
    img_dim = 2 if backend.image_data_format() == 'channels_first' else 1
    input_size = backend.int_shape(inputs)[img_dim:(img_dim + 2)]

    if isinstance(kernel_size, int):
        kernel_size = (kernel_size, kernel_size)

    if input_size[0] is None:
        adjust = (1, 1)
    else:
        adjust = (1 - input_size[0] % 2, 1 - input_size[1] % 2)

    correct = (kernel_size[0] // 2, kernel_size[1] // 2)

    return ((correct[0] - adjust[0], correct[0]),
            (correct[1] - adjust[1], correct[1])) 

def get_swish(**kwargs):
    def swish(x):
        """Swish activation function: x * sigmoid(x).
        Reference: [Searching for Activation Functions](https://arxiv.org/abs/1710.05941)
        """

        if backend.backend() == 'tensorflow':
            try:
                # The native TF implementation has a more
                # memory-efficient gradient implementation
                return backend.tf.nn.swish(x)
            except AttributeError:
                pass

        return x * backend.sigmoid(x)
    return swish 

def __call__(cls, *args, **kwargs):
        obj = cls.__new__(cls, *args, **kwargs)
        from .keras_model import KerasModel
        if issubclass(cls, KerasModel):
            from tensorflow.keras import backend as K
            if K.backend() != 'tensorflow':
                obj.__init__(*args, **kwargs)
                return obj

            K.clear_session()
            obj.graph = tf.Graph()
            with obj.graph.as_default():
                if hasattr(cls, '_config_session'):
                    obj.sess = cls._config_session()
                else:
                    obj.sess = tf.Session()
        else:
            obj.graph = tf.Graph()

        for meth in dir(obj):
            if meth == '__class__':
                continue
            attr = getattr(obj, meth)
            if callable(attr):
                if issubclass(cls, KerasModel):
                    wrapped_attr = _keras_wrap(attr, obj.sess)
                else:
                    wrapped_attr = _graph_wrap(attr, obj.graph)
                setattr(obj, meth, wrapped_attr)
        obj.__init__(*args, **kwargs)
        return obj 

def _keras_wrap(func, session):
    """Constructs function encapsulated in the graph and the session."""
    @wraps(func)
    def _wrapped(*args, **kwargs):
        with session.graph.as_default():
            tf.keras.backend.set_session(session)
            return func(*args, **kwargs)

    return _wrapped 

def tpu_compatible():
    '''Fit the tpu problems we meet while using keras tpu model'''
    if not hasattr(tpu_compatible, 'once'):
        tpu_compatible.once = True
    else:
        return
    import tensorflow as tf
    import tensorflow.keras.backend as K
    _version = tf.__version__.split('.')
    is_correct_version = int(_version[0]) >= 1 and (int(_version[0]) >= 2 or int(_version[1]) >= 13)
    from tensorflow.contrib.tpu.python.tpu.keras_support import KerasTPUModel
    def initialize_uninitialized_variables():
        sess = K.get_session()
        uninitialized_variables = set([i.decode('ascii') for i in sess.run(tf.report_uninitialized_variables())])
        init_op = tf.variables_initializer(
            [v for v in tf.global_variables() if v.name.split(':')[0] in uninitialized_variables]
        )
        sess.run(init_op)

    _tpu_compile = KerasTPUModel.compile

    def tpu_compile(self,
                    optimizer,
                    loss=None,
                    metrics=None,
                    loss_weights=None,
                    sample_weight_mode=None,
                    weighted_metrics=None,
                    target_tensors=None,
                    **kwargs):
        if not is_correct_version:
            raise ValueError('You need tensorflow >= 1.3 for better keras tpu support!')
        _tpu_compile(self, optimizer, loss, metrics, loss_weights,
                     sample_weight_mode, weighted_metrics,
                     target_tensors, **kwargs)
        initialize_uninitialized_variables()  # for unknown reason, we should run this after compile sometimes

    KerasTPUModel.compile = tpu_compile 

def test_replace_imports():
    python_code = """
    import keras
    from keras import backend as K
    import os
    import keras_contrib
    import keras_contrib.layers as lay
    import keras.layers
    from keras.layers import Dense

    if K.backend() == 'tensorflow':
        import tensorflow as tf
        function = tf.max
    """

    expected_code = """
    from tensorflow import keras
    from tensorflow.keras import backend as K
    import os
    import keras_contrib
    import keras_contrib.layers as lay
    import tensorflow.keras.layers
    from tensorflow.keras.layers import Dense

    if K.backend() == 'tensorflow':
        import tensorflow as tf
        function = tf.max
    """

    code_with_replacement = replace_imports_in_text(python_code, False)
    assert expected_code == code_with_replacement
    assert python_code == replace_imports_in_text(code_with_replacement, True) 

def preprocess_input(x):
    """
    "mode" option description in preprocess_input
    mode: One of "caffe", "tf" or "torch".
        - caffe: will convert the images from RGB to BGR,
            then will zero-center each color channel with
            respect to the ImageNet dataset,
            without scaling.
        - tf: will scale pixels between -1 and 1,
            sample-wise.
        - torch: will scale pixels between 0 and 1 and then
            will normalize each channel with respect to the
            ImageNet dataset.
    """
    x = _preprocess_input(x, mode='tf', backend=K)
    #x /= 255.
    #mean = [0.485, 0.456, 0.406]
    #std = [0.229, 0.224, 0.225]

    #x[..., 0] -= mean[0]
    #x[..., 1] -= mean[1]
    #x[..., 2] -= mean[2]
    #if std is not None:
        #x[..., 0] /= std[0]
        #x[..., 1] /= std[1]
        #x[..., 2] /= std[2]

    return x 

def clear_session_after_test():
    """Test wrapper to clean up after TensorFlow and CNTK tests.

    This wrapper runs for all the tests in the keras test suite.
    """
    yield
    if K.backend() == 'tensorflow' or K.backend() == 'cntk':
        K.clear_session() 

def refresh_keras_backend(use_tpu=True):
    clean_keras_module()
    import keras.backend as K
    if use_tpu and K.backend() != 'theano':
        clean_keras_module()
        replace_keras_to_tf_keras()
        import keras.backend as K
    return K 

def replace_keras_to_tf_keras():
    tpu_compatible()
    import tensorflow as tf
    sys.modules['keras'] = tf.keras
    globals()['keras'] = tf.keras
    import keras.backend as K
    K.tf = tf 

def block1(x, filters, kernel_size=3, stride=1,
           conv_shortcut=True, name=None):
    """A residual block.
    # Arguments
        x: input tensor.
        filters: integer, filters of the bottleneck layer.
        kernel_size: default 3, kernel size of the bottleneck layer.
        stride: default 1, stride of the first layer.
        conv_shortcut: default True, use convolution shortcut if True,
            otherwise identity shortcut.
        name: string, block label.
    # Returns
        Output tensor for the residual block.
    """
    bn_axis = 3 if backend.image_data_format() == 'channels_last' else 1

    if conv_shortcut is True:
        shortcut = layers.Conv2D(4 * filters, 1, strides=stride,
                                 name=name + '_0_conv')(x)
        shortcut = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                             name=name + '_0_bn')(shortcut)
    else:
        shortcut = x

    x = layers.Conv2D(filters, 1, strides=stride, name=name + '_1_conv')(x)
    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                  name=name + '_1_bn')(x)
    x = layers.Activation('relu', name=name + '_1_relu')(x)

    x = layers.Conv2D(filters, kernel_size, padding='same',
                      name=name + '_2_conv')(x)
    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                  name=name + '_2_bn')(x)
    x = layers.Activation('relu', name=name + '_2_relu')(x)

    x = layers.Conv2D(4 * filters, 1, name=name + '_3_conv')(x)
    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                  name=name + '_3_bn')(x)

    x = layers.Add(name=name + '_add')([shortcut, x])
    x = layers.Activation('relu', name=name + '_out')(x)
    return x 

def block2(x, filters, kernel_size=3, stride=1,
           conv_shortcut=False, name=None):
    """A residual block.
    # Arguments
        x: input tensor.
        filters: integer, filters of the bottleneck layer.
        kernel_size: default 3, kernel size of the bottleneck layer.
        stride: default 1, stride of the first layer.
        conv_shortcut: default False, use convolution shortcut if True,
            otherwise identity shortcut.
        name: string, block label.
    # Returns
        Output tensor for the residual block.
    """
    bn_axis = 3 if backend.image_data_format() == 'channels_last' else 1

    preact = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                       name=name + '_preact_bn')(x)
    preact = layers.Activation('relu', name=name + '_preact_relu')(preact)

    if conv_shortcut is True:
        shortcut = layers.Conv2D(4 * filters, 1, strides=stride,
                                 name=name + '_0_conv')(preact)
    else:
        shortcut = layers.MaxPooling2D(
            1, strides=stride)(x) if stride > 1 else x

    x = layers.Conv2D(filters, 1, strides=1, use_bias=False,
                      name=name + '_1_conv')(preact)
    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                  name=name + '_1_bn')(x)
    x = layers.Activation('relu', name=name + '_1_relu')(x)

    x = layers.ZeroPadding2D(padding=((1, 1), (1, 1)), name=name + '_2_pad')(x)
    x = layers.Conv2D(filters, kernel_size, strides=stride,
                      use_bias=False, name=name + '_2_conv')(x)
    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                  name=name + '_2_bn')(x)
    x = layers.Activation('relu', name=name + '_2_relu')(x)

    x = layers.Conv2D(4 * filters, 1, name=name + '_3_conv')(x)
    x = layers.Add(name=name + '_out')([shortcut, x])
    return x 

def block3(x, filters, kernel_size=3, stride=1, groups=32,
           conv_shortcut=True, name=None):
    """A residual block.
    # Arguments
        x: input tensor.
        filters: integer, filters of the bottleneck layer.
        kernel_size: default 3, kernel size of the bottleneck layer.
        stride: default 1, stride of the first layer.
        groups: default 32, group size for grouped convolution.
        conv_shortcut: default True, use convolution shortcut if True,
            otherwise identity shortcut.
        name: string, block label.
    # Returns
        Output tensor for the residual block.
    """
    bn_axis = 3 if backend.image_data_format() == 'channels_last' else 1

    if conv_shortcut is True:
        shortcut = layers.Conv2D((64 // groups) * filters, 1, strides=stride,
                                 use_bias=False, name=name + '_0_conv')(x)
        shortcut = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                             name=name + '_0_bn')(shortcut)
    else:
        shortcut = x

    x = layers.Conv2D(filters, 1, use_bias=False, name=name + '_1_conv')(x)
    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                  name=name + '_1_bn')(x)
    x = layers.Activation('relu', name=name + '_1_relu')(x)

    c = filters // groups
    x = layers.ZeroPadding2D(padding=((1, 1), (1, 1)), name=name + '_2_pad')(x)
    x = layers.DepthwiseConv2D(kernel_size, strides=stride, depth_multiplier=c,
                               use_bias=False, name=name + '_2_conv')(x)
    x_shape = backend.int_shape(x)[1:-1]
    x = layers.Reshape(x_shape + (groups, c, c))(x)
    output_shape = x_shape + (groups,
                              c) if backend.backend() == 'theano' else None

    x = layers.Lambda(lambda x: sum([x[:, :, :, :, i] for i in range(c)]),
                      output_shape=output_shape, name=name + '_2_reduce')(x)

    x = layers.Reshape(x_shape + (filters,))(x)

    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                  name=name + '_2_bn')(x)

    x = layers.Activation('relu', name=name + '_2_relu')(x)

    x = layers.Conv2D((64 // groups) * filters, 1, use_bias=False,
                      name=name + '_3_conv')(x)

    x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-5,
                                  name=name + '_3_bn')(x)

    x = layers.Add(name=name + '_add')([shortcut, x])
    x = layers.Activation('relu', name=name + '_out')(x)
    return x