Python tensorflow.contrib.framework.python.ops.arg_scope() Examples

def vgg_arg_scope(weight_decay=0.0005):
  """Defines the VGG arg scope.

  Args:
    weight_decay: The l2 regularization coefficient.

  Returns:
    An arg_scope.
  """
  with arg_scope(
      [layers.conv2d, layers_lib.fully_connected],
      activation_fn=nn_ops.relu,
      weights_regularizer=regularizers.l2_regularizer(weight_decay),
      biases_initializer=init_ops.zeros_initializer()):
    with arg_scope([layers.conv2d], padding='SAME') as arg_sc:
      return arg_sc 

def testFullyConvolutionalEndpointShapes(self):
    global_pool = False
    num_classes = 10
    inputs = create_test_input(2, 321, 321, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_small(
          inputs, num_classes, global_pool, scope='resnet')
      endpoint_to_shape = {
          'resnet/block1': [2, 41, 41, 4],
          'resnet/block2': [2, 21, 21, 8],
          'resnet/block3': [2, 11, 11, 16],
          'resnet/block4': [2, 11, 11, 32]
      }
      for endpoint in endpoint_to_shape:
        shape = endpoint_to_shape[endpoint]
        self.assertListEqual(end_points[endpoint].get_shape().as_list(), shape) 

def testClassificationShapes(self):
    global_pool = True
    num_classes = 10
    inputs = create_test_input(2, 224, 224, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_small(
          inputs, num_classes, global_pool, scope='resnet')
      endpoint_to_shape = {
          'resnet/block1': [2, 28, 28, 4],
          'resnet/block2': [2, 14, 14, 8],
          'resnet/block3': [2, 7, 7, 16],
          'resnet/block4': [2, 7, 7, 32]
      }
      for endpoint in endpoint_to_shape:
        shape = endpoint_to_shape[endpoint]
        self.assertListEqual(end_points[endpoint].get_shape().as_list(), shape) 

def testAtrousFullyConvolutionalValues(self):
    """Verify dense feature extraction with atrous convolution."""
    nominal_stride = 32
    for output_stride in [4, 8, 16, 32, None]:
      with arg_scope(resnet_utils.resnet_arg_scope(is_training=False)):
        with ops.Graph().as_default():
          with self.test_session() as sess:
            random_seed.set_random_seed(0)
            inputs = create_test_input(2, 81, 81, 3)
            # Dense feature extraction followed by subsampling.
            output, _ = self._resnet_small(
                inputs, None, global_pool=False, output_stride=output_stride)
            if output_stride is None:
              factor = 1
            else:
              factor = nominal_stride // output_stride
            output = resnet_utils.subsample(output, factor)
            # Make the two networks use the same weights.
            variable_scope.get_variable_scope().reuse_variables()
            # Feature extraction at the nominal network rate.
            expected, _ = self._resnet_small(inputs, None, global_pool=False)
            sess.run(variables.global_variables_initializer())
            self.assertAllClose(
                output.eval(), expected.eval(), atol=1e-4, rtol=1e-4) 

def testRootlessFullyConvolutionalEndpointShapes(self):
    global_pool = False
    num_classes = 10
    inputs = create_test_input(2, 128, 128, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_small(
          inputs,
          num_classes,
          global_pool,
          include_root_block=False,
          scope='resnet')
      endpoint_to_shape = {
          'resnet/block1': [2, 64, 64, 4],
          'resnet/block2': [2, 32, 32, 8],
          'resnet/block3': [2, 16, 16, 16],
          'resnet/block4': [2, 16, 16, 32]
      }
      for endpoint in endpoint_to_shape:
        shape = endpoint_to_shape[endpoint]
        self.assertListEqual(end_points[endpoint].get_shape().as_list(), shape) 

def testAtrousFullyConvolutionalEndpointShapes(self):
    global_pool = False
    num_classes = 10
    output_stride = 8
    inputs = create_test_input(2, 321, 321, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_small(
          inputs,
          num_classes,
          global_pool,
          output_stride=output_stride,
          scope='resnet')
      endpoint_to_shape = {
          'resnet/block1': [2, 41, 41, 4],
          'resnet/block2': [2, 41, 41, 8],
          'resnet/block3': [2, 41, 41, 16],
          'resnet/block4': [2, 41, 41, 32]
      }
      for endpoint in endpoint_to_shape:
        shape = endpoint_to_shape[endpoint]
        self.assertListEqual(end_points[endpoint].get_shape().as_list(), shape) 

def testAtrousFullyConvolutionalValues(self):
    """Verify dense feature extraction with atrous convolution."""
    nominal_stride = 32
    for output_stride in [4, 8, 16, 32, None]:
      with arg_scope(resnet_utils.resnet_arg_scope(is_training=False)):
        with ops.Graph().as_default():
          with self.test_session() as sess:
            random_seed.set_random_seed(0)
            inputs = create_test_input(2, 81, 81, 3)
            # Dense feature extraction followed by subsampling.
            output, _ = self._resnet_small(
                inputs, None, global_pool=False, output_stride=output_stride)
            if output_stride is None:
              factor = 1
            else:
              factor = nominal_stride // output_stride
            output = resnet_utils.subsample(output, factor)
            # Make the two networks use the same weights.
            variable_scope.get_variable_scope().reuse_variables()
            # Feature extraction at the nominal network rate.
            expected, _ = self._resnet_small(inputs, None, global_pool=False)
            sess.run(variables.global_variables_initializer())
            self.assertAllClose(
                output.eval(), expected.eval(), atol=1e-4, rtol=1e-4) 

def testAtrousFullyConvolutionalEndpointShapes(self):
    global_pool = False
    num_classes = 10
    output_stride = 8
    inputs = create_test_input(2, 321, 321, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_small(
          inputs,
          num_classes,
          global_pool,
          output_stride=output_stride,
          scope='resnet')
      endpoint_to_shape = {
          'resnet/block1': [2, 41, 41, 4],
          'resnet/block2': [2, 41, 41, 8],
          'resnet/block3': [2, 41, 41, 16],
          'resnet/block4': [2, 41, 41, 32]
      }
      for endpoint in endpoint_to_shape:
        shape = endpoint_to_shape[endpoint]
        self.assertListEqual(end_points[endpoint].get_shape().as_list(), shape) 

def testRootlessFullyConvolutionalEndpointShapes(self):
    global_pool = False
    num_classes = 10
    inputs = create_test_input(2, 128, 128, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_small(
          inputs,
          num_classes,
          global_pool,
          include_root_block=False,
          scope='resnet')
      endpoint_to_shape = {
          'resnet/block1': [2, 64, 64, 4],
          'resnet/block2': [2, 32, 32, 8],
          'resnet/block3': [2, 16, 16, 16],
          'resnet/block4': [2, 16, 16, 32]
      }
      for endpoint in endpoint_to_shape:
        shape = endpoint_to_shape[endpoint]
        self.assertListEqual(end_points[endpoint].get_shape().as_list(), shape) 

def testFullyConvolutionalEndpointShapes(self):
    global_pool = False
    num_classes = 10
    inputs = create_test_input(2, 321, 321, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_small(
          inputs, num_classes, global_pool, scope='resnet')
      endpoint_to_shape = {
          'resnet/block1': [2, 41, 41, 4],
          'resnet/block2': [2, 21, 21, 8],
          'resnet/block3': [2, 11, 11, 16],
          'resnet/block4': [2, 11, 11, 32]
      }
      for endpoint in endpoint_to_shape:
        shape = endpoint_to_shape[endpoint]
        self.assertListEqual(end_points[endpoint].get_shape().as_list(), shape) 

def vgg_arg_scope(weight_decay=0.0005):
  """Defines the VGG arg scope.

  Args:
    weight_decay: The l2 regularization coefficient.

  Returns:
    An arg_scope.
  """
  with arg_scope(
      [layers.conv2d, layers_lib.fully_connected],
      activation_fn=nn_ops.relu,
      weights_regularizer=regularizers.l2_regularizer(weight_decay),
      biases_initializer=init_ops.zeros_initializer()):
    with arg_scope([layers.conv2d], padding='SAME') as arg_sc:
      return arg_sc 

def up(self, input, **_):
        hps = self.hps
        h_size = hps.h_size
        z_size = hps.z_size
        stride = [2, 2] if self.downsample else [1, 1]

        with arg_scope([conv2d]):
            x = tf.nn.elu(input)
            x = conv2d("up_conv1", x, 2 * z_size + 2 * h_size, stride=stride)
            self.qz_mean, self.qz_logsd, self.up_context, h = split(x, 1, [z_size, z_size, h_size, h_size])

            h = tf.nn.elu(h)
            h = conv2d("up_conv3", h, h_size)
            if self.downsample:
                input = resize_nearest_neighbor(input, 0.5)
            return input + 0.1 * h 

def vgg_arg_scope(weight_decay=0.0005):
    """Defines the VGG arg scope.

    Args:
      weight_decay: The l2 regularization coefficient.

    Returns:
      An arg_scope.
    """
    with arg_scope(
        [layers.conv2d, layers_lib.fully_connected],
        activation_fn=nn_ops.relu,
        weights_regularizer=regularizers.l2_regularizer(weight_decay),
        biases_initializer=init_ops.zeros_initializer()
    ):
        with arg_scope([layers.conv2d], padding='SAME') as arg_sc:
            return arg_sc 

def vgg_arg_scope(weight_decay=0.0005):
    """Defines the VGG arg scope.

    Args:
      weight_decay: The l2 regularization coefficient.

    Returns:
      An arg_scope.
    """
    with arg_scope(
        [layers.conv2d, layers_lib.fully_connected],
        activation_fn=nn_ops.relu,
        weights_regularizer=regularizers.l2_regularizer(weight_decay),
        biases_initializer=init_ops.zeros_initializer()
    ):
        with arg_scope([layers.conv2d], padding='SAME') as arg_sc:
            return arg_sc 

def vgg_arg_scope(weight_decay=0.0005):
    """Defines the VGG arg scope.

    Args:
      weight_decay: The l2 regularization coefficient.

    Returns:
      An arg_scope.
    """
    with arg_scope(
        [layers.conv2d, layers_lib.fully_connected],
        activation_fn=nn_ops.relu,
        weights_regularizer=regularizers.l2_regularizer(weight_decay),
        biases_initializer=init_ops.zeros_initializer()
    ):
        with arg_scope([layers.conv2d], padding='SAME') as arg_sc:
            return arg_sc 

def predictron_arg_scope(weight_decay=0.0001,
                         batch_norm_decay=0.997,
                         batch_norm_epsilon=1e-5,
                         batch_norm_scale=True):
  batch_norm_params = {
    'decay': batch_norm_decay,
    'epsilon': batch_norm_epsilon,
    'scale': batch_norm_scale,
    'updates_collections': tf.GraphKeys.UPDATE_OPS,
  }

  # Set weight_decay for weights in Conv and FC layers.
  with arg_scope(
      [layers.conv2d, layers_lib.fully_connected],
      weights_regularizer=regularizers.l2_regularizer(weight_decay)):
    with arg_scope(
        [layers.conv2d],
        weights_initializer=initializers.variance_scaling_initializer(),
        activation_fn=None,
        normalizer_fn=layers_lib.batch_norm,
        normalizer_params=batch_norm_params) as sc:
      return sc 

def vgg_arg_scope(weight_decay=0.0005):
  """Defines the VGG arg scope.

  Args:
    weight_decay: The l2 regularization coefficient.

  Returns:
    An arg_scope.
  """
  with arg_scope(
      [layers.conv2d, layers_lib.fully_connected],
      activation_fn=nn_ops.relu,
      weights_regularizer=regularizers.l2_regularizer(weight_decay),
      biases_initializer=init_ops.zeros_initializer()):
    with arg_scope([layers.conv2d], padding='SAME') as arg_sc:
      return arg_sc 

def testClassificationShapes(self):
    global_pool = True
    num_classes = 10
    inputs = create_test_input(2, 224, 224, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_small(
          inputs, num_classes, global_pool, scope='resnet')
      endpoint_to_shape = {
          'resnet/block1': [2, 28, 28, 4],
          'resnet/block2': [2, 14, 14, 8],
          'resnet/block3': [2, 7, 7, 16],
          'resnet/block4': [2, 7, 7, 32]
      }
      for endpoint in endpoint_to_shape:
        shape = endpoint_to_shape[endpoint]
        self.assertListEqual(end_points[endpoint].get_shape().as_list(), shape) 

def actnorm(name, x, scale=1., logdet=None, logscale_factor=3., batch_variance=False, reverse=False, init=False, trainable=True):
    if arg_scope([get_variable_ddi], trainable=trainable):
        if not reverse:
            x = actnorm_center(name+"_center", x, reverse)
            x = actnorm_scale(name+"_scale", x, scale, logdet,
                              logscale_factor, batch_variance, reverse, init)
            if logdet != None:
                x, logdet = x
        else:
            x = actnorm_scale(name + "_scale", x, scale, logdet,
                              logscale_factor, batch_variance, reverse, init)
            if logdet != None:
                x, logdet = x
            x = actnorm_center(name+"_center", x, reverse)
        if logdet != None:
            return x, logdet
        return x

# Activation normalization 

def testEndPointsV1(self):
    """Test the end points of a tiny v1 bottleneck network."""
    bottleneck = resnet_v1.bottleneck
    blocks = [
        resnet_utils.Block('block1', bottleneck, [(4, 1, 1), (4, 1, 2)]),
        resnet_utils.Block('block2', bottleneck, [(8, 2, 1), (8, 2, 1)])
    ]
    inputs = create_test_input(2, 32, 16, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      _, end_points = self._resnet_plain(inputs, blocks, scope='tiny')
    expected = [
        'tiny/block1/unit_1/bottleneck_v1/shortcut',
        'tiny/block1/unit_1/bottleneck_v1/shortcut/BatchNorm',
        'tiny/block1/unit_1/bottleneck_v1/conv1',
        'tiny/block1/unit_1/bottleneck_v1/conv2',
        'tiny/block1/unit_1/bottleneck_v1/conv3',
        'tiny/block1/unit_1/bottleneck_v1/conv3/BatchNorm',
        'tiny/block1/unit_2/bottleneck_v1/conv1',
        'tiny/block1/unit_2/bottleneck_v1/conv2',
        'tiny/block1/unit_2/bottleneck_v1/conv3',
        'tiny/block1/unit_2/bottleneck_v1/conv3/BatchNorm',
        'tiny/block2/unit_1/bottleneck_v1/shortcut',
        'tiny/block2/unit_1/bottleneck_v1/shortcut/BatchNorm',
        'tiny/block2/unit_1/bottleneck_v1/conv1',
        'tiny/block2/unit_1/bottleneck_v1/conv2',
        'tiny/block2/unit_1/bottleneck_v1/conv3',
        'tiny/block2/unit_1/bottleneck_v1/conv3/BatchNorm',
        'tiny/block2/unit_2/bottleneck_v1/conv1',
        'tiny/block2/unit_2/bottleneck_v1/conv2',
        'tiny/block2/unit_2/bottleneck_v1/conv3',
        'tiny/block2/unit_2/bottleneck_v1/conv3/BatchNorm'
    ]
    self.assertItemsEqual(expected, end_points) 

def testFullyConvolutionalUnknownHeightWidth(self):
    batch = 2
    height, width = 65, 65
    global_pool = False
    inputs = create_test_input(batch, None, None, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      output, _ = self._resnet_small(inputs, None, global_pool)
    self.assertListEqual(output.get_shape().as_list(), [batch, None, None, 32])
    images = create_test_input(batch, height, width, 3)
    with self.test_session() as sess:
      sess.run(variables.global_variables_initializer())
      output = sess.run(output, {inputs: images.eval()})
      self.assertEqual(output.shape, (batch, 3, 3, 32)) 

def _resnet_plain(self, inputs, blocks, output_stride=None, scope=None):
    """A plain ResNet without extra layers before or after the ResNet blocks."""
    with variable_scope.variable_scope(scope, values=[inputs]):
      with arg_scope([layers.conv2d], outputs_collections='end_points'):
        net = resnet_utils.stack_blocks_dense(inputs, blocks, output_stride)
        end_points = utils.convert_collection_to_dict('end_points')
        return net, end_points 

def testModelHasExpectedNumberOfParameters(self):
    batch_size = 5
    height, width = 299, 299
    inputs = random_ops.random_uniform((batch_size, height, width, 3))
    with arg_scope(inception_v3.inception_v3_arg_scope()):
      inception_v3.inception_v3_base(inputs)
    total_params, _ = model_analyzer.analyze_vars(
        variables_lib.get_model_variables())
    self.assertAlmostEqual(21802784, total_params) 

def __call__(self, features):
    """ Define tf graph.
    """
    inputs = features['image']

    with tf.variable_scope('encoder') as vsc:
      with slim.arg_scope(resnet_v2.resnet_arg_scope()):
        # conv1
        with arg_scope(
            [layers_lib.conv2d], activation_fn=None, normalizer_fn=None):
          net = resnet_utils.conv2d_same(inputs, 16, 5, stride=2, scope='conv1')
        tf.add_to_collection(vsc.original_name_scope, net)

        # resnet blocks
        blocks = []
        for i in range(len(self.encoder_params['block_name'])):
          block = resnet_v2.resnet_v2_block(
              scope=self.encoder_params['block_name'][i],
              base_depth=self.encoder_params['base_depth'][i],
              num_units=self.encoder_params['num_units'][i],
              stride=self.encoder_params['stride'][i])
          blocks.append(block)
        net, _ = resnet_v2.resnet_v2(
            net,
            blocks,
            is_training=(self.mode == ModeKeys.TRAIN),
            global_pool=False,
            output_stride=2,
            include_root_block=False,
            scope='resnet')

        tf.add_to_collection(vsc.original_name_scope, net)
    return net 

def ar_multiconv2d(name, x, context, n_h, n_out, nl=tf.nn.elu, **_):
    with tf.variable_scope(name), arg_scope([ar_conv2d]):
        for i, size in enumerate(n_h):
            x = ar_conv2d("layer_%d" % i, x, size, zerodiagonal=False)
            if i == 0:
                x += context
            x = nl(x)
        return [ar_conv2d("layer_out_%d" % i, x, size, zerodiagonal=True) for i, size in enumerate(n_out)] 

def ar_conv2d(name, x, num_filters, filter_size=(3, 3), stride=(1, 1), pad="SAME", init_scale=1.,
              zerodiagonal=True, **_):
    h = filter_size[0]
    w = filter_size[1]
    n_in = int(x.get_shape()[1])
    n_out = num_filters

    mask = tf.constant(get_conv_ar_mask(h, w, n_in, n_out, zerodiagonal))
    with arg_scope([conv2d]):
        return conv2d(name, x, num_filters, filter_size, stride, pad, init_scale, mask=mask)


# Auto-Regressive convnet with l2 normalization 

def testModelHasExpectedNumberOfParameters(self):
    batch_size = 5
    height, width = 224, 224
    inputs = random_ops.random_uniform((batch_size, height, width, 3))
    with arg_scope(inception_v1.inception_v1_arg_scope()):
      inception_v1.inception_v1_base(inputs)
    total_params, _ = model_analyzer.analyze_vars(
        variables_lib.get_model_variables())
    self.assertAlmostEqual(5607184, total_params) 

def alexnet_v2_arg_scope(weight_decay=0.0005):
  with arg_scope(
      [layers.conv2d, layers_lib.fully_connected],
      activation_fn=nn_ops.relu,
      biases_initializer=init_ops.constant_initializer(0.1),
      weights_regularizer=regularizers.l2_regularizer(weight_decay)):
    with arg_scope([layers.conv2d], padding='SAME'):
      with arg_scope([layers_lib.max_pool2d], padding='VALID') as arg_sc:
        return arg_sc 

def testAtrousFullyConvolutionalUnknownHeightWidth(self):
    batch = 2
    height, width = 65, 65
    global_pool = False
    output_stride = 8
    inputs = create_test_input(batch, None, None, 3)
    with arg_scope(resnet_utils.resnet_arg_scope()):
      output, _ = self._resnet_small(
          inputs, None, global_pool, output_stride=output_stride)
    self.assertListEqual(output.get_shape().as_list(), [batch, None, None, 32])
    images = create_test_input(batch, height, width, 3)
    with self.test_session() as sess:
      sess.run(variables.global_variables_initializer())
      output = sess.run(output, {inputs: images.eval()})
      self.assertEqual(output.shape, (batch, 9, 9, 32)) 

def overfeat_arg_scope(weight_decay=0.0005):
  with arg_scope(
      [layers.conv2d, layers_lib.fully_connected],
      activation_fn=nn_ops.relu,
      weights_regularizer=regularizers.l2_regularizer(weight_decay),
      biases_initializer=init_ops.zeros_initializer()):
    with arg_scope([layers.conv2d], padding='SAME'):
      with arg_scope([layers_lib.max_pool2d], padding='VALID') as arg_sc:
        return arg_sc