Python layers.conv2d函数代码示例

def conv2d_same(inputs, num_outputs, kernel_size, stride, rate=1, scope=None):
  """Strided 2-D convolution with 'SAME' padding.

  When stride > 1, then we do explicit zero-padding, followed by conv2d with
  'VALID' padding.

  Note that

     net = conv2d_same(inputs, num_outputs, 3, stride=stride)

  is equivalent to

     net = tf.contrib.layers.conv2d(inputs, num_outputs, 3, stride=1,
     padding='SAME')
     net = subsample(net, factor=stride)

  whereas

     net = tf.contrib.layers.conv2d(inputs, num_outputs, 3, stride=stride,
     padding='SAME')

  is different when the input's height or width is even, which is why we add the
  current function. For more details, see ResnetUtilsTest.testConv2DSameEven().

  Args:
    inputs: A 4-D tensor of size [batch, height_in, width_in, channels].
    num_outputs: An integer, the number of output filters.
    kernel_size: An int with the kernel_size of the filters.
    stride: An integer, the output stride.
    rate: An integer, rate for atrous convolution.
    scope: Scope.

  Returns:
    output: A 4-D tensor of size [batch, height_out, width_out, channels] with
      the convolution output.
  """
  if stride == 1:
    return layers_lib.conv2d(
        inputs,
        num_outputs,
        kernel_size,
        stride=1,
        rate=rate,
        padding='SAME',
        scope=scope)
  else:
    kernel_size_effective = kernel_size + (kernel_size - 1) * (rate - 1)
    pad_total = kernel_size_effective - 1
    pad_beg = pad_total // 2
    pad_end = pad_total - pad_beg
    inputs = array_ops.pad(
        inputs, [[0, 0], [pad_beg, pad_end], [pad_beg, pad_end], [0, 0]])
    return layers_lib.conv2d(
        inputs,
        num_outputs,
        kernel_size,
        stride=stride,
        rate=rate,
        padding='VALID',
        scope=scope)

def conv_model(X, Y_, mode):
    XX = tf.reshape(X, [-1, 28, 28, 1])
    biasInit = tf.constant_initializer(0.1, dtype=tf.float32)
    Y1 = layers.conv2d(XX,  num_outputs=6,  kernel_size=[6, 6], biases_initializer=biasInit)
    Y2 = layers.conv2d(Y1, num_outputs=12, kernel_size=[5, 5], stride=2, biases_initializer=biasInit)
    Y3 = layers.conv2d(Y2, num_outputs=24, kernel_size=[4, 4], stride=2, biases_initializer=biasInit)
    Y4 = layers.flatten(Y3)
    Y5 = layers.relu(Y4, 200, biases_initializer=biasInit)
    # to deactivate dropout on the dense layer, set keep_prob=1
    Y5d = layers.dropout(Y5, keep_prob=0.75, noise_shape=None, is_training=mode==learn.ModeKeys.TRAIN)
    Ylogits = layers.linear(Y5d, 10)
    predict = tf.nn.softmax(Ylogits)
    classes = tf.cast(tf.argmax(predict, 1), tf.uint8)

    loss = conv_model_loss(Ylogits, Y_, mode)
    train_op = conv_model_train_op(loss, mode)
    eval_metrics = conv_model_eval_metrics(classes, Y_, mode)

    return learn.ModelFnOps(
        mode=mode,
        # You can name the fields of your predictions dictionary as you like.
        predictions={"predictions": predict, "classes": classes},
        loss=loss,
        train_op=train_op,
        eval_metric_ops=eval_metrics
    )

    def discriminator_stego_nn(self, img, reuse=False):
        with tf.variable_scope('S_network'):

            if reuse:
                tf.get_variable_scope().reuse_variables()

            net = img
            net = self.image_processing_layer(img)
            net = self.batch_norm(net, scope='d_s_bn0')
            net = conv2d(net, self.df_dim, kernel_size=[5, 5], stride=[2, 2],
                         activation_fn=self.leaky_relu, scope='d_s_h0_conv')

            net = self.batch_norm(net, scope='d_s_bn1')
            net = conv2d(net, self.df_dim * 2, kernel_size=[5, 5], stride=[2, 2],
                         activation_fn=self.leaky_relu, scope='d_s_h1_conv')

            net = self.batch_norm(net, scope='d_s_bn2')
            net = conv2d(net, self.df_dim * 4, kernel_size=[5, 5], stride=[2, 2],
                         activation_fn=self.leaky_relu, scope='d_s_h2_conv')

            net = self.batch_norm(net, scope='d_s_bn3')
            net = conv2d(net, self.df_dim * 8, kernel_size=[5, 5], stride=[2, 2],
                         activation_fn=self.leaky_relu, scope='d_s_h3_conv')

            net = self.batch_norm(net, scope='d_s_bn4')

            net = tf.reshape(net, [self.conf.batch_size, -1])
            net = linear(net, 1, activation_fn=tf.nn.sigmoid, scope='d_s_h4_lin',
                         weights_initializer=tf.random_normal_initializer(stddev=0.02))

            return net

def vgg_a(inputs,
          num_classes=1000,
          is_training=True,
          dropout_keep_prob=0.5,
          spatial_squeeze=True,
          scope='vgg_a'):
  """Oxford Net VGG 11-Layers version A Example.

  Note: All the fully_connected layers have been transformed to conv2d layers.
        To use in classification mode, resize input to 224x224.

  Args:
    inputs: a tensor of size [batch_size, height, width, channels].
    num_classes: number of predicted classes.
    is_training: whether or not the model is being trained.
    dropout_keep_prob: the probability that activations are kept in the dropout
      layers during training.
    spatial_squeeze: whether or not should squeeze the spatial dimensions of the
      outputs. Useful to remove unnecessary dimensions for classification.
    scope: Optional scope for the variables.

  Returns:
    the last op containing the log predictions and end_points dict.
  """
  with variable_scope.variable_scope(scope, 'vgg_a', [inputs]) as sc:
    end_points_collection = sc.original_name_scope + '_end_points'
    # Collect outputs for conv2d, fully_connected and max_pool2d.
    with arg_scope(
        [layers.conv2d, layers_lib.max_pool2d],
        outputs_collections=end_points_collection):
      net = layers_lib.repeat(
          inputs, 1, layers.conv2d, 64, [3, 3], scope='conv1')
      net = layers_lib.max_pool2d(net, [2, 2], scope='pool1')
      net = layers_lib.repeat(net, 1, layers.conv2d, 128, [3, 3], scope='conv2')
      net = layers_lib.max_pool2d(net, [2, 2], scope='pool2')
      net = layers_lib.repeat(net, 2, layers.conv2d, 256, [3, 3], scope='conv3')
      net = layers_lib.max_pool2d(net, [2, 2], scope='pool3')
      net = layers_lib.repeat(net, 2, layers.conv2d, 512, [3, 3], scope='conv4')
      net = layers_lib.max_pool2d(net, [2, 2], scope='pool4')
      net = layers_lib.repeat(net, 2, layers.conv2d, 512, [3, 3], scope='conv5')
      net = layers_lib.max_pool2d(net, [2, 2], scope='pool5')
      # Use conv2d instead of fully_connected layers.
      net = layers.conv2d(net, 4096, [7, 7], padding='VALID', scope='fc6')
      net = layers_lib.dropout(
          net, dropout_keep_prob, is_training=is_training, scope='dropout6')
      net = layers.conv2d(net, 4096, [1, 1], scope='fc7')
      net = layers_lib.dropout(
          net, dropout_keep_prob, is_training=is_training, scope='dropout7')
      net = layers.conv2d(
          net,
          num_classes, [1, 1],
          activation_fn=None,
          normalizer_fn=None,
          scope='fc8')
      # Convert end_points_collection into a end_point dict.
      end_points = utils.convert_collection_to_dict(end_points_collection)
      if spatial_squeeze:
        net = array_ops.squeeze(net, [1, 2], name='fc8/squeezed')
        end_points[sc.name + '/fc8'] = net
      return net, end_points

def bottleneck(inputs,
               depth,
               depth_bottleneck,
               stride,
               rate=1,
               outputs_collections=None,
               scope=None):
  """Bottleneck residual unit variant with BN before convolutions.

  This is the full preactivation residual unit variant proposed in [2]. See
  Fig. 1(b) of [2] for its definition. Note that we use here the bottleneck
  variant which has an extra bottleneck layer.

  When putting together two consecutive ResNet blocks that use this unit, one
  should use stride = 2 in the last unit of the first block.

  Args:
    inputs: A tensor of size [batch, height, width, channels].
    depth: The depth of the ResNet unit output.
    depth_bottleneck: The depth of the bottleneck layers.
    stride: The ResNet unit's stride. Determines the amount of downsampling of
      the units output compared to its input.
    rate: An integer, rate for atrous convolution.
    outputs_collections: Collection to add the ResNet unit output.
    scope: Optional variable_scope.

  Returns:
    The ResNet unit's output.
  """
  with variable_scope.variable_scope(scope, 'bottleneck_v2', [inputs]) as sc:
    depth_in = utils.last_dimension(inputs.get_shape(), min_rank=4)
    preact = layers.batch_norm(
        inputs, activation_fn=nn_ops.relu, scope='preact')
    if depth == depth_in:
      shortcut = resnet_utils.subsample(inputs, stride, 'shortcut')
    else:
      shortcut = layers_lib.conv2d(
          preact,
          depth, [1, 1],
          stride=stride,
          normalizer_fn=None,
          activation_fn=None,
          scope='shortcut')

    residual = layers_lib.conv2d(
        preact, depth_bottleneck, [1, 1], stride=1, scope='conv1')
    residual = resnet_utils.conv2d_same(
        residual, depth_bottleneck, 3, stride, rate=rate, scope='conv2')
    residual = layers_lib.conv2d(
        residual,
        depth, [1, 1],
        stride=1,
        normalizer_fn=None,
        activation_fn=None,
        scope='conv3')

    output = shortcut + residual

    return utils.collect_named_outputs(outputs_collections, sc.name, output)

def build_atari(minimap, screen, info, msize, ssize, num_action):
  # Extract features
  mconv1 = layers.conv2d(tf.transpose(minimap, [0, 2, 3, 1]),
                         num_outputs=16,
                         kernel_size=8,
                         stride=4,
                         scope='mconv1')
  mconv2 = layers.conv2d(mconv1,
                         num_outputs=32,
                         kernel_size=4,
                         stride=2,
                         scope='mconv2')
  sconv1 = layers.conv2d(tf.transpose(screen, [0, 2, 3, 1]),
                         num_outputs=16,
                         kernel_size=8,
                         stride=4,
                         scope='sconv1')
  sconv2 = layers.conv2d(sconv1,
                         num_outputs=32,
                         kernel_size=4,
                         stride=2,
                         scope='sconv2')
  info_fc = layers.fully_connected(layers.flatten(info),
                                   num_outputs=256,
                                   activation_fn=tf.tanh,
                                   scope='info_fc')

  # Compute spatial actions, non spatial actions and value
  feat_fc = tf.concat([layers.flatten(mconv2), layers.flatten(sconv2), info_fc], axis=1)
  feat_fc = layers.fully_connected(feat_fc,
                                   num_outputs=256,
                                   activation_fn=tf.nn.relu,
                                   scope='feat_fc')

  spatial_action_x = layers.fully_connected(feat_fc,
                                            num_outputs=ssize,
                                            activation_fn=tf.nn.softmax,
                                            scope='spatial_action_x')
  spatial_action_y = layers.fully_connected(feat_fc,
                                            num_outputs=ssize,
                                            activation_fn=tf.nn.softmax,
                                            scope='spatial_action_y')
  spatial_action_x = tf.reshape(spatial_action_x, [-1, 1, ssize])
  spatial_action_x = tf.tile(spatial_action_x, [1, ssize, 1])
  spatial_action_y = tf.reshape(spatial_action_y, [-1, ssize, 1])
  spatial_action_y = tf.tile(spatial_action_y, [1, 1, ssize])
  spatial_action = layers.flatten(spatial_action_x * spatial_action_y)

  non_spatial_action = layers.fully_connected(feat_fc,
                                              num_outputs=num_action,
                                              activation_fn=tf.nn.softmax,
                                              scope='non_spatial_action')
  value = tf.reshape(layers.fully_connected(feat_fc,
                                            num_outputs=1,
                                            activation_fn=None,
                                            scope='value'), [-1])

  return spatial_action, non_spatial_action, value

def conv_model(X, Y_):
   XX = tf.reshape(X, [-1, 28, 28, 1])
   Y1 = layers.conv2d(XX,  num_outputs=6,  kernel_size=[6, 6])
   Y2 = layers.conv2d(Y1,  num_outputs=12, kernel_size=[5, 5], stride=2)
   Y3 = layers.conv2d(Y2,  num_outputs=24, kernel_size=[4, 4], stride=2)
   Y4 = layers.flatten(Y3)
   Y5 = layers.relu(Y4, 200)
   Ylogits = layers.linear(Y5, 10)
   predict = tf.nn.softmax(Ylogits)

   classes = tf.cast(tf.argmax(predict, 1), tf.uint8)
   loss = tf.nn.softmax_cross_entropy_with_logits(Ylogits, tf.one_hot(Y_, 10))
   train_op = layers.optimize_loss(loss, framework.get_global_step(), 0.003, "Adam")
   return {"predictions":predict, "classes": classes}, loss, train_op

def conv_model(X, Y_, mode):
    XX = tf.reshape(X, [-1, 28, 28, 1])
    biasInit = tf.constant_initializer(0.1, dtype=tf.float32)
    Y1 = layers.conv2d(XX,  num_outputs=6,  kernel_size=[6, 6], biases_initializer=biasInit)
    Y2 = layers.conv2d(Y1, num_outputs=12, kernel_size=[5, 5], stride=2, biases_initializer=biasInit)
    Y3 = layers.conv2d(Y2, num_outputs=24, kernel_size=[4, 4], stride=2, biases_initializer=biasInit)
    Y4 = layers.flatten(Y3)
    Y5 = layers.relu(Y4, 200, biases_initializer=biasInit)
    Ylogits = layers.linear(Y5, 10)
    predict = tf.nn.softmax(Ylogits)
    classes = tf.cast(tf.argmax(predict, 1), tf.uint8)
    loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(Ylogits, tf.one_hot(Y_, 10)))*100
    train_op = layers.optimize_loss(loss, framework.get_global_step(), 0.001, "Adam")
    return {"predictions":predict, "classes": classes}, loss, train_op

def resBlock(x, num_outputs, kernel_size = 4, stride=1, activation_fn=tf.nn.relu, normalizer_fn=tcl.batch_norm, scope=None):
    assert num_outputs%2==0 #num_outputs must be divided by channel_factor(2 here)
    with tf.variable_scope(scope, 'resBlock'):
        shortcut = x
        if stride != 1 or x.get_shape()[3] != num_outputs:
            shortcut = tcl.conv2d(shortcut, num_outputs, kernel_size=1, stride=stride, 
                        activation_fn=None, normalizer_fn=None, scope='shortcut')
        x = tcl.conv2d(x, num_outputs/2, kernel_size=1, stride=1, padding='SAME')
        x = tcl.conv2d(x, num_outputs/2, kernel_size=kernel_size, stride=stride, padding='SAME')
        x = tcl.conv2d(x, num_outputs, kernel_size=1, stride=1, activation_fn=None, padding='SAME', normalizer_fn=None)

        x += shortcut       
        x = normalizer_fn(x)
        x = activation_fn(x)
    return x

	def __call__(self, x, reuse=False):
		with tf.variable_scope(self.name) as scope:
			if reuse:
				scope.reuse_variables()
			size = 64
			shared = tcl.conv2d(x, num_outputs=size, kernel_size=5, # bzx28x28x1 -> bzx14x14x64
						stride=2, activation_fn=tf.nn.relu)
			shared = tcl.conv2d(shared, num_outputs=size * 2, kernel_size=5, # 7x7x128
						stride=2, activation_fn=lrelu, normalizer_fn=tcl.batch_norm)
			shared = tcl.fully_connected(tcl.flatten( # reshape, 1
						shared), 1024, activation_fn=tf.nn.relu)
			
			#c = tcl.fully_connected(shared, 128, activation_fn=tf.nn.relu, normalizer_fn=tcl.batch_norm)
			c = tcl.fully_connected(shared, 10, activation_fn=None) # 10 classes
			return c

	def __call__(self, x, reuse=True):
		with tf.variable_scope(self.name) as scope:
			if reuse:
				scope.reuse_variables()
			size = 64
			d = tcl.conv2d(x, num_outputs=size, kernel_size=3, # bzx64x64x3 -> bzx32x32x64
						stride=2, activation_fn=lrelu)
			d = tcl.conv2d(d, num_outputs=size * 2, kernel_size=3, # 16x16x128
						stride=2, activation_fn=lrelu, normalizer_fn=tcl.batch_norm)
			d = tcl.conv2d(d, num_outputs=size * 4, kernel_size=3, # 8x8x256
						stride=2, activation_fn=lrelu, normalizer_fn=tcl.batch_norm)
			d = tcl.conv2d(d, num_outputs=size * 8, kernel_size=3, # 4x4x512
						stride=2, activation_fn=lrelu, normalizer_fn=tcl.batch_norm)
			d = tcl.fully_connected(tcl.flatten( # reshape, 1
						d), 1, activation_fn=None)
			return d

    def get_bottlenet(self):
        """Form the network"""
        # get collections
        bottlelayer = namedtuple("bottlelayer",
                              ['kernel_shape','stride','bn_flag','padding','act_fn'])
        with tf.name_scope(self.scope):
            input_now = self.inputs
            for i, kernel in enumerate(self.bottle_params):
                with tf.name_scope('bottle_sub'+str('i')):
                    kernel = bottlelayer._make(kernel)
                    with tf.name_scope('conv2d'):
                        residual = conv2d(
                            inputs=input_now,
                            num_outputs=kernel.kernel_shape[-1],
                            kernel_size=kernel.kernel_shape[0:2],
                            padding=kernel.padding,
                            stride=kernel.stride,
                            )
                    if kernel.bn_flag:
                        residual = utils.get_batch_norm(residual,
                                                        self.is_training,
                                                        scope='batch_norm')
                    if kernel.act_fn is not None:
                        with tf.name_scope('activate'):
                            residual = kernel.act_fn(residual)
                    input_now = residual
            # add shortcut
            self.get_shortcut(self.stride,scope=self.scope+'_shortcut')
            residual = residual + self.shortcut
            if self.summary_flag:
                tf.summary.histogram('bottle_residual', residual)

        return residual

def autoencoder(inputs):
    # encoder
    # 32 x 32 x 1   ->  16 x 16 x 32
    # 16 x 16 x 32  ->  8 x 8 x 16
    # 8 x 8 x 16    ->  2 x 2 x 8
    net = lays.conv2d(inputs, 32, [5, 5], stride=2, padding='SAME')
    net = lays.conv2d(net, 16, [5, 5], stride=2, padding='SAME')
    net = lays.conv2d(net, 8, [5, 5], stride=4, padding='SAME')
    # decoder
    # 2 x 2 x 8    ->  8 x 8 x 16
    # 8 x 8 x 16   ->  16 x 16 x 32
    # 16 x 16 x 32  ->  32 x 32 x 1
    net = lays.conv2d_transpose(net, 16, [5, 5], stride=4, padding='SAME')
    net = lays.conv2d_transpose(net, 32, [5, 5], stride=2, padding='SAME')
    net = lays.conv2d_transpose(net, 1, [5, 5], stride=2, padding='SAME', activation_fn=tf.nn.tanh)
    return net

    def get_shortcut(self, stride, scope='shortcut'):
        """Reshape and repeat to get the shortcut of input

        Reference
        =========
        [1] TensorFlow 实战
        """
        def subsample(inputs, factor, scope):
            if factor == 1:
                return inputs
            else:
                # avg for auto encoder
                return avg_pool2d(inputs,[1,1],
                                  stride=factor,
                                  padding='SAME',
                                  scope=scope)
        if self.depth_in == self.depth_out:
            self.shortcut = subsample(self.inputs, stride, scope)
        else:
            self.shortcut = conv2d(
                inputs=self.inputs,
                num_outputs=self.depth_out,
                kernel_size=[1,1],
                stride=stride,
                padding='SAME',
                normalizer_fn=None,
                activation_fn=None,
                scope=scope)

def conv2d(input, filters=16, kernel_size=3, padding="same", stride=1, activation_fn=relu, reuse=None, name=None, data_format='NHWC',
           weights_initializer=xavier_initializer(), biases_initializer=tf.zeros_initializer(), weights_regularizer=None, biases_regularizer=None,
           normalizer_fn=None, normalizer_params=None):

   return contrib_layers.conv2d(input, num_outputs=filters, kernel_size=kernel_size, padding=padding, stride=stride, scope=name, data_format=data_format,
                        activation_fn=activation_fn, reuse=reuse, weights_initializer=weights_initializer, biases_initializer=biases_initializer,
                        weights_regularizer=weights_regularizer, biases_regularizer=biases_regularizer, normalizer_fn=normalizer_fn, normalizer_params=normalizer_params)