python类reshape()的实例源码

def conv(input, kernel, biases, k_h, k_w, c_o, s_h, s_w,  padding="VALID", group=1):
    '''From https://github.com/ethereon/caffe-tensorflow
    '''
    c_i = input.get_shape()[-1]
    assert c_i%group==0
    assert c_o%group==0
    convolve = lambda i, k: tf.nn.conv2d(i, k, [1, s_h, s_w, 1], padding=padding)


    if group==1:
        conv = convolve(input, kernel)
    else:
        input_groups = tf.split(3, group, input)
        kernel_groups = tf.split(3, group, kernel)
        output_groups = [convolve(i, k) for i,k in zip(input_groups, kernel_groups)]
        conv = tf.concat(3, output_groups)
    return  tf.reshape(tf.nn.bias_add(conv, biases), [-1]+conv.get_shape().as_list()[1:])

def dense(inputs, units, bias_shape, w_i, b_i=None, activation=tf.nn.relu):
    # ??tf.layers?????flatten
    # dense1 = tf.layers.dense(tf.contrib.layers.flatten(relu5), activation=tf.nn.relu, units=50)
    if not isinstance(inputs, ops.Tensor):
        inputs = ops.convert_to_tensor(inputs, dtype='float')
        # dim_list = inputs.get_shape().as_list()
        # flatten_shape = dim_list[1] if len(dim_list) <= 2 else reduce(lambda x, y: x * y, dim_list[1:])
        # reshaped = tf.reshape(inputs, [dim_list[0], flatten_shape])
    if len(inputs.shape) > 2:
        inputs = tf.contrib.layers.flatten(inputs)
    flatten_shape = inputs.shape[1]
    weights = tf.get_variable('weights', shape=[flatten_shape, units], initializer=w_i)
    dense = tf.matmul(inputs, weights)
    if bias_shape is not None:
        assert bias_shape[0] == units
        biases = tf.get_variable('biases', shape=bias_shape, initializer=b_i)
        return activation(dense + biases) if activation is not None else dense + biases
    return activation(dense) if activation is not None else dense

def conv(input, kernel, biases, k_h, k_w, c_o, s_h, s_w, padding="VALID", group=1):
    '''From https://github.com/ethereon/caffe-tensorflow
    '''
    c_i = input.get_shape()[-1]
    assert c_i % group == 0
    assert c_o % group == 0
    convolve = lambda i, k: tf.nn.conv2d(i, k, [1, s_h, s_w, 1], padding=padding)

    if group == 1:
        conv = convolve(input, kernel)
    else:
        input_groups = tf.split(3, group, input)
        kernel_groups = tf.split(3, group, kernel)
        output_groups = [convolve(i, k) for i, k in zip(input_groups, kernel_groups)]
        conv = tf.concat(3, output_groups)
    return tf.reshape(tf.nn.bias_add(conv, biases), [-1] + conv.get_shape().as_list()[1:])

def conv(input, kernel, biases, k_h, k_w, c_o, s_h, s_w, padding="VALID", group=1):
    '''From https://github.com/ethereon/caffe-tensorflow
    '''
    c_i = input.get_shape()[-1]
    assert c_i % group == 0
    assert c_o % group == 0
    convolve = lambda i, k: tf.nn.conv2d(i, k, [1, s_h, s_w, 1], padding=padding)

    if group == 1:
        conv = convolve(input, kernel)
    else:
        input_groups = tf.split(3, group, input)
        kernel_groups = tf.split(3, group, kernel)
        output_groups = [convolve(i, k) for i, k in zip(input_groups, kernel_groups)]
        conv = tf.concat(3, output_groups)
    return tf.reshape(tf.nn.bias_add(conv, biases), [-1] + conv.get_shape().as_list()[1:])

def conv(input, kernel, biases, k_h, k_w, c_o, s_h, s_w, padding="VALID", group=1):
    '''From https://github.com/ethereon/caffe-tensorflow
    '''
    c_i = input.get_shape()[-1]
    assert c_i % group == 0
    assert c_o % group == 0
    convolve = lambda i, k: tf.nn.conv2d(i, k, [1, s_h, s_w, 1], padding=padding)

    if group == 1:
        conv = convolve(input, kernel)
    else:
        input_groups = tf.split(3, group, input)
        kernel_groups = tf.split(3, group, kernel)
        output_groups = [convolve(i, k) for i, k in zip(input_groups, kernel_groups)]
        conv = tf.concat(3, output_groups)
    return tf.reshape(tf.nn.bias_add(conv, biases), [-1] + conv.get_shape().as_list()[1:])

def calculate_loss_distill_relabel(self, predictions, labels_distill, labels, **unused_params):
    with tf.name_scope("loss_distill_relabel"):
      print("loss_distill_relabel")
      epsilon = 10e-6
      float_labels = tf.cast(labels, tf.float32)
      sum_labels = tf.cast(tf.reduce_sum(float_labels),dtype=tf.int32)
      pos_distill, _ = tf.nn.top_k(tf.reshape(labels_distill,[-1]), k=sum_labels)
      labels_true = tf.ones(tf.shape(labels))
      labels_false = tf.zeros(tf.shape(labels))
      labels_add = tf.where(tf.greater_equal(labels_distill, pos_distill[-1]), labels_true, labels_false)
      print(labels_add.get_shape().as_list())
      float_labels = float_labels+labels_add*(1.0-float_labels)
      cross_entropy_loss = float_labels * tf.log(predictions + epsilon) + (
          1 - float_labels) * tf.log(1 - predictions + epsilon)
      cross_entropy_loss = tf.negative(cross_entropy_loss)

      return tf.reduce_mean(tf.reduce_sum(cross_entropy_loss, 1))

def get_forward_parameters(vocab_size=4716):
    t_vars = tf.trainable_variables()
    h1_vars_weight = [var for var in t_vars if 'hidden_1' in var.name and 'weights' in var.name]
    h1_vars_biases = [var for var in t_vars if 'hidden_1' in var.name and 'biases' in var.name]
    h2_vars_weight = [var for var in t_vars if 'hidden_2' in var.name and 'weights' in var.name]
    h2_vars_biases = [var for var in t_vars if 'hidden_2' in var.name and 'biases' in var.name]
    o1_vars_weight = [var for var in t_vars if 'output_1' in var.name and 'weights' in var.name]
    o1_vars_biases = [var for var in t_vars if 'output_1' in var.name and 'biases' in var.name]
    o2_vars_weight = [var for var in t_vars if 'output_2' in var.name and 'weights' in var.name]
    o2_vars_biases = [var for var in t_vars if 'output_2' in var.name and 'biases' in var.name]
    h1_vars_biases = tf.reshape(h1_vars_biases[0],[1,FLAGS.hidden_size_1])
    h2_vars_biases = tf.reshape(h2_vars_biases[0],[1,FLAGS.hidden_size_2])
    o1_vars_biases = tf.reshape(o1_vars_biases[0],[1,FLAGS.hidden_size_1])
    o2_vars_biases = tf.reshape(o2_vars_biases[0],[1,vocab_size])
    vars_1 = tf.concat((h1_vars_weight[0],h1_vars_biases),axis=0)
    vars_2 = tf.concat((h2_vars_weight[0],h2_vars_biases),axis=0)
    vars_3 = tf.concat((o1_vars_weight[0],o1_vars_biases),axis=0)
    vars_4 = tf.concat((o2_vars_weight[0],o2_vars_biases),axis=0)
    return [vars_1,vars_2,vars_3,vars_4]

def get_forward_parameters(vocab_size=4716):
    t_vars = tf.trainable_variables()
    h1_vars_weight = [var for var in t_vars if 'hidden_1' in var.name and 'weights' in var.name]
    h1_vars_biases = [var for var in t_vars if 'hidden_1' in var.name and 'biases' in var.name]
    h2_vars_weight = [var for var in t_vars if 'hidden_2' in var.name and 'weights' in var.name]
    h2_vars_biases = [var for var in t_vars if 'hidden_2' in var.name and 'biases' in var.name]
    o1_vars_weight = [var for var in t_vars if 'output_1' in var.name and 'weights' in var.name]
    o1_vars_biases = [var for var in t_vars if 'output_1' in var.name and 'biases' in var.name]
    o2_vars_weight = [var for var in t_vars if 'output_2' in var.name and 'weights' in var.name]
    o2_vars_biases = [var for var in t_vars if 'output_2' in var.name and 'biases' in var.name]
    h1_vars_biases = tf.reshape(h1_vars_biases[0],[1,FLAGS.hidden_size_1])
    h2_vars_biases = tf.reshape(h2_vars_biases[0],[1,FLAGS.hidden_size_2])
    o1_vars_biases = tf.reshape(o1_vars_biases[0],[1,FLAGS.hidden_size_1])
    o2_vars_biases = tf.reshape(o2_vars_biases[0],[1,vocab_size])
    vars_1 = tf.concat((h1_vars_weight[0],h1_vars_biases),axis=0)
    vars_2 = tf.concat((h2_vars_weight[0],h2_vars_biases),axis=0)
    vars_3 = tf.concat((o1_vars_weight[0],o1_vars_biases),axis=0)
    vars_4 = tf.concat((o2_vars_weight[0],o2_vars_biases),axis=0)
    return [vars_1,vars_2,vars_3,vars_4]

def FramePooling(frames, method, **unused_params):
  """Pools over the frames of a video.

  Args:
    frames: A tensor with shape [batch_size, num_frames, feature_size].
    method: "average", "max", "attention", or "none".
  Returns:
    A tensor with shape [batch_size, feature_size] for average, max, or
    attention pooling. A tensor with shape [batch_size*num_frames, feature_size]
    for none pooling.

  Raises:
    ValueError: if method is other than "average", "max", "attention", or
    "none".
  """
  if method == "average":
    return tf.reduce_mean(frames, 1)
  elif method == "max":
    return tf.reduce_max(frames, 1)
  elif method == "none":
    feature_size = frames.shape_as_list()[2]
    return tf.reshape(frames, [-1, feature_size])
  else:
    raise ValueError("Unrecognized pooling method: %s" % method)

def calculate_loss(self, predictions, support_predictions, labels, **unused_params):
    """ 
    support_predictions batch_size x num_models x num_classes
    predictions = tf.reduce_mean(support_predictions, axis=1)
    """
    model_count = tf.shape(support_predictions)[1]
    vocab_size = tf.shape(support_predictions)[2]

    mean_predictions = tf.reduce_mean(support_predictions, axis=1, keep_dims=True)
    support_labels = tf.tile(tf.expand_dims(tf.cast(labels, dtype=tf.float32), axis=1), multiples=[1,model_count,1])
    support_means = tf.stop_gradient(tf.tile(mean_predictions, multiples=[1,model_count,1]))

    support_predictions = tf.reshape(support_predictions, shape=[-1,model_count*vocab_size])
    support_labels = tf.reshape(support_labels, shape=[-1,model_count*vocab_size])
    support_means = tf.reshape(support_means, shape=[-1,model_count*vocab_size])

    ce_loss_fn = CrossEntropyLoss()
    # The cross entropy between predictions and ground truth
    cross_entropy_loss = ce_loss_fn.calculate_loss(support_predictions, support_labels, **unused_params)
    # The cross entropy between predictions and mean predictions
    divergence = ce_loss_fn.calculate_loss(support_predictions, support_means, **unused_params)

    loss = cross_entropy_loss * (1.0 - FLAGS.support_loss_percent) - divergence * FLAGS.support_loss_percent
    return loss

def resolution(self, model_input_raw, num_frames, resolution, method="SELECT"):
    frame_dim = len(model_input_raw.get_shape()) - 2
    feature_dim = len(model_input_raw.get_shape()) - 1
    max_frames = model_input_raw.get_shape().as_list()[frame_dim]
    num_features = model_input_raw.get_shape().as_list()[feature_dim]
    if resolution > 1:
      new_max_frames = max_frames / resolution
      cut_frames = new_max_frames * resolution
      model_input_raw = model_input_raw[:, :cut_frames, :]
      model_input_raw = tf.reshape(model_input_raw, shape=[-1,new_max_frames,resolution,num_features])
      if method == "MEAN":
        model_input_raw = tf.reduce_mean(model_input_raw, axis=2)
      elif method == "MAX":
        model_input_raw = tf.reduce_max(model_input_raw, axis=2)
      elif method == "SELECT":
        model_input_raw = model_input_raw[:,:,resolution-1,:]
      model_input = tf.nn.l2_normalize(model_input_raw, feature_dim)
      num_frames = num_frames / resolution
    else:
      model_input = tf.nn.l2_normalize(model_input_raw, feature_dim)
    return model_input, num_frames

def resolution(self, model_input_raw, num_frames, resolution, method="SELECT"):
    frame_dim = len(model_input_raw.get_shape()) - 2
    feature_dim = len(model_input_raw.get_shape()) - 1
    max_frames = model_input_raw.get_shape().as_list()[frame_dim]
    num_features = model_input_raw.get_shape().as_list()[feature_dim]
    if resolution > 1:
      new_max_frames = max_frames / resolution
      cut_frames = new_max_frames * resolution
      model_input_raw = model_input_raw[:, :cut_frames, :]
      model_input_raw = tf.reshape(model_input_raw, shape=[-1,new_max_frames,resolution,num_features])
      if method == "MEAN":
        model_input_raw = tf.reduce_mean(model_input_raw, axis=2)
      elif method == "MAX":
        model_input_raw = tf.reduce_max(model_input_raw, axis=2)
      elif method == "SELECT":
        model_input_raw = model_input_raw[:,:,resolution-1,:]
      model_input = tf.nn.l2_normalize(model_input_raw, feature_dim)
      num_frames = num_frames / resolution
    else:
      model_input = tf.nn.l2_normalize(model_input_raw, feature_dim)
    return model_input, num_frames

def resolution(self, model_input_raw, num_frames, resolution):
    frame_dim = len(model_input_raw.get_shape()) - 2
    feature_dim = len(model_input_raw.get_shape()) - 1
    max_frames = model_input_raw.get_shape().as_list()[frame_dim]
    num_features = model_input_raw.get_shape().as_list()[feature_dim]
    if resolution > 1:
      new_max_frames = max_frames / resolution
      cut_frames = new_max_frames * resolution
      model_input_raw = model_input_raw[:, :cut_frames, :]
      model_input_raw = tf.reshape(model_input_raw, shape=[-1,new_max_frames,resolution,num_features])
      model_input_raw = tf.reduce_mean(model_input_raw, axis=2)

      model_input = tf.nn.l2_normalize(model_input_raw, feature_dim)
      num_frames = num_frames / resolution
    else:
      model_input = tf.nn.l2_normalize(model_input_raw, feature_dim)
    return model_input, num_frames

def resolution(self, model_input_raw, num_frames):
    resolution = FLAGS.time_resolution

    frame_dim = len(model_input_raw.get_shape()) - 2
    feature_dim = len(model_input_raw.get_shape()) - 1

    max_frames = model_input_raw.get_shape().as_list()[frame_dim]
    num_features = model_input_raw.get_shape().as_list()[feature_dim]

    new_max_frames = max_frames / resolution
    cut_frames = new_max_frames * resolution

    model_input_raw = model_input_raw[:, :cut_frames, :]
    model_input_raw = tf.reshape(model_input_raw, shape=[-1,new_max_frames,resolution,num_features])
    model_input_raw = tf.reduce_mean(model_input_raw, axis=2)
    num_frames = num_frames / resolution
    return model_input_raw, num_frames

def FramePooling(frames, method, **unused_params):
  """Pools over the frames of a video.

  Args:
    frames: A tensor with shape [batch_size, num_frames, feature_size].
    method: "average", "max", "attention", or "none".
  Returns:
    A tensor with shape [batch_size, feature_size] for average, max, or
    attention pooling. A tensor with shape [batch_size*num_frames, feature_size]
    for none pooling.

  Raises:
    ValueError: if method is other than "average", "max", "attention", or
    "none".
  """
  if method == "average":
    return tf.reduce_mean(frames, 1)
  elif method == "max":
    return tf.reduce_max(frames, 1)
  elif method == "none":
    feature_size = frames.shape_as_list()[2]
    return tf.reshape(frames, [-1, feature_size])
  else:
    raise ValueError("Unrecognized pooling method: %s" % method)

def FramePooling(frames, method, **unused_params):
  """Pools over the frames of a video.

  Args:
    frames: A tensor with shape [batch_size, num_frames, feature_size].
    method: "average", "max", "attention", or "none".
  Returns:
    A tensor with shape [batch_size, feature_size] for average, max, or
    attention pooling. A tensor with shape [batch_size*num_frames, feature_size]
    for none pooling.

  Raises:
    ValueError: if method is other than "average", "max", "attention", or
    "none".
  """
  if method == "average":
    return tf.reduce_mean(frames, 1)
  elif method == "max":
    return tf.reduce_max(frames, 1)
  elif method == "none":
    feature_size = frames.shape_as_list()[2]
    return tf.reshape(frames, [-1, feature_size])
  else:
    raise ValueError("Unrecognized pooling method: %s" % method)

def __init__(self, ob_space, ac_space, size=256, **kwargs):
        self.x = x = tf.placeholder(tf.float32, [None] + list(ob_space))

        for i in range(4):
            x = tf.nn.elu(conv2d(x, 32, "l{}".format(i + 1), [3, 3], [2, 2]))
        # introduce a "fake" batch dimension of 1 after flatten so that we can do GRU over time dim
        x = tf.expand_dims(flatten(x), 1)

        gru = rnn.GRUCell(size)

        h_init = np.zeros((1, size), np.float32)
        self.state_init = [h_init]
        h_in = tf.placeholder(tf.float32, [1, size])
        self.state_in = [h_in]

        gru_outputs, gru_state = tf.nn.dynamic_rnn(
            gru, x, initial_state=h_in, sequence_length=[size], time_major=True)
        x = tf.reshape(gru_outputs, [-1, size])
        self.logits = linear(x, ac_space, "action", normalized_columns_initializer(0.01))
        self.vf = tf.reshape(linear(x, 1, "value", normalized_columns_initializer(1.0)), [-1])
        self.state_out = [gru_state[:1]]
        self.sample = categorical_sample(self.logits, ac_space)[0, :]
        self.var_list = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, tf.get_variable_scope().name)

def __init__(self, ob_space, ac_space, layers=[256], **kwargs):
        self.x = x = tf.placeholder(tf.float32, [None] + list(ob_space))

        rank = len(ob_space)

        if rank == 3: # pixel input
            for i in range(4):
                x = tf.nn.elu(conv2d(x, 32, "c{}".format(i + 1), [3, 3], [2, 2]))
        elif rank == 1: # plain features
            #x = tf.nn.elu(linear(x, 256, "l1", normalized_columns_initializer(0.01)))
            pass
        else:
            raise TypeError("observation space must have rank 1 or 3, got %d" % rank)

        x = flatten(x)

        for i, layer in enumerate(layers):
            x = tf.nn.elu(linear(x, layer, "l{}".format(i + 1), tf.contrib.layers.xavier_initializer()))

        self.logits = linear(x, ac_space, "action", tf.contrib.layers.xavier_initializer())
        self.vf = tf.reshape(linear(x, 1, "value", tf.contrib.layers.xavier_initializer()), [-1])
        self.sample = categorical_sample(self.logits, ac_space)[0, :]
        self.var_list = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, tf.get_variable_scope().name)
        self.state_in = []

def __init__(self, ob_space, ac_space, size=256, **kwargs):
        self.x = x = tf.placeholder(tf.float32, [None] + list(ob_space))

        for i in range(4):
            x = tf.nn.elu(conv2d(x, 32, "l{}".format(i + 1), [3, 3], [2, 2]))
        # introduce a "fake" batch dimension of 1 after flatten so that we can do GRU over time dim
        x = tf.expand_dims(flatten(x), 1)

        gru = rnn.GRUCell(size)

        h_init = np.zeros((1, size), np.float32)
        self.state_init = [h_init]
        h_in = tf.placeholder(tf.float32, [1, size])
        self.state_in = [h_in]

        gru_outputs, gru_state = tf.nn.dynamic_rnn(
            gru, x, initial_state=h_in, sequence_length=[size], time_major=True)
        x = tf.reshape(gru_outputs, [-1, size])
        self.logits = linear(x, ac_space, "action", normalized_columns_initializer(0.01))
        self.vf = tf.reshape(linear(x, 1, "value", normalized_columns_initializer(1.0)), [-1])
        self.state_out = [gru_state[:1]]
        self.sample = categorical_sample(self.logits, ac_space)[0, :]
        self.var_list = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, tf.get_variable_scope().name)

def baseline_forward(self, X, size, n_class):
        shape = X.get_shape()
        _X = tf.transpose(X, [1, 0, 2])  # batch_size x sentence_length x word_length -> batch_size x sentence_length x word_length
        _X = tf.reshape(_X, [-1, int(shape[2])])  # (batch_size x sentence_length) x word_length
        seq = tf.split(0, int(shape[1]), _X)  # sentence_length x (batch_size x word_length)

        with tf.name_scope("LSTM"):
            lstm_cell = rnn_cell.BasicLSTMCell(size, forget_bias=1.0)
            outputs, states = rnn.rnn(lstm_cell, seq, dtype=tf.float32)

        with tf.name_scope("LSTM-Classifier"):
            W = tf.Variable(tf.random_normal([size, n_class]), name="W")
            b = tf.Variable(tf.random_normal([n_class]), name="b")
            output = tf.matmul(outputs[-1], W) + b

        return output

def discriminate(self, x_var, y, weights, biases, reuse=False):

        y1 =  tf.reshape(y, shape=[self.batch_size, 1, 1, self.y_dim])
        x_var = conv_cond_concat(x_var, y1)

        conv1= lrelu(conv2d(x_var, weights['wc1'], biases['bc1']))

        conv1 = conv_cond_concat(conv1, y1)

        conv2= lrelu(batch_normal(conv2d(conv1, weights['wc2'], biases['bc2']), scope='dis_bn1', reuse=reuse))

        conv2 = tf.reshape(conv2, [self.batch_size, -1])

        conv2 = tf.concat([conv2, y], 1)

        fc1 = lrelu(batch_normal(fully_connect(conv2, weights['wc3'], biases['bc3']), scope='dis_bn2', reuse=reuse))

        fc1 = tf.concat([fc1, y], 1)
        #for D
        output= fully_connect(fc1, weights['wd'], biases['bd'])

        return tf.nn.sigmoid(output)

def encode_z(self, x, weights, biases):

        c1 = tf.nn.relu(batch_normal(conv2d(x, weights['e1'], biases['eb1']), scope='enz_bn1'))

        c2 = tf.nn.relu(batch_normal(conv2d(c1, weights['e2'], biases['eb2']), scope='enz_bn2'))

        c2 = tf.reshape(c2, [self.batch_size, 128*7*7])

        #using tanh instead of tf.nn.relu.
        result_z = batch_normal(fully_connect(c2, weights['e3'], biases['eb3']), scope='enz_bn3')

        #result_c = tf.nn.sigmoid(fully_connect(c2, weights['e4'], biases['eb4']))

        #Transforming one-hot form
        #sparse_label = tf.arg_max(result_c, 1)

        #y_vec = tf.one_hot(sparse_label, 10)

        return result_z

def _create(self):
    # Concat bridge inputs on the depth dimensions
    bridge_input = nest.map_structure(
        lambda x: tf.reshape(x, [self.batch_size, _total_tensor_depth(x)]),
        self._bridge_input)
    bridge_input_flat = nest.flatten([bridge_input])
    bridge_input_concat = tf.concat(bridge_input_flat, 1)

    state_size_splits = nest.flatten(self.decoder_state_size)
    total_decoder_state_size = sum(state_size_splits)

    # Pass bridge inputs through a fully connected layer layer
    initial_state_flat = tf.contrib.layers.fully_connected(
        inputs=bridge_input_concat,
        num_outputs=total_decoder_state_size,
        activation_fn=self._activation_fn)

    # Shape back into required state size
    initial_state = tf.split(initial_state_flat, state_size_splits, axis=1)
    return nest.pack_sequence_as(self.decoder_state_size, initial_state)

def encode(self, inputs):
    inputs = tf.image.resize_images(
        images=inputs,
        size=[self.params["resize_height"], self.params["resize_width"]],
        method=tf.image.ResizeMethod.BILINEAR)

    outputs, _ = inception_v3_base(tf.to_float(inputs))
    output_shape = outputs.get_shape()  #pylint: disable=E1101
    shape_list = output_shape.as_list()

    # Take attentin over output elemnts in width and height dimension:
    # Shape: [B, W*H, ...]
    outputs_flat = tf.reshape(outputs, [shape_list[0], -1, shape_list[-1]])

    # Final state is the pooled output
    # Shape: [B, W*H*...]
    final_state = tf.contrib.slim.avg_pool2d(
        outputs, output_shape[1:3], padding="VALID", scope="pool")
    final_state = tf.contrib.slim.flatten(outputs, scope="flatten")

    return EncoderOutput(
        outputs=outputs_flat,
        final_state=final_state,
        attention_values=outputs_flat,
        attention_values_length=tf.shape(outputs_flat)[1])

def get_training_tensors(self, learning_rate = 0.001, grad_clip = 5):
        #-----------------------------------------------------------------------
        # Build a loss function
        #-----------------------------------------------------------------------
        with tf.name_scope('targets-encode'):
            y_one_hot  = tf.one_hot(self.targets, self.n_classes)
            y_reshaped = tf.reshape(y_one_hot, self.logits.get_shape())

        with tf.name_scope('loss'):
            loss = tf.nn.softmax_cross_entropy_with_logits(logits=self.logits,
                                                           labels=y_reshaped)
            loss = tf.reduce_mean(loss)
            tf.summary.scalar('loss', loss)

        #-----------------------------------------------------------------------
        # Build the optimizer
        #-----------------------------------------------------------------------
        with tf.name_scope('optimizer'):
            tvars     = tf.trainable_variables()
            grads, _  = tf.clip_by_global_norm(tf.gradients(loss, tvars),
                                               grad_clip)
            train_op  = tf.train.AdamOptimizer(learning_rate)
            optimizer = train_op.apply_gradients(zip(grads, tvars))

        return loss, optimizer

def plotFields(layer,fieldShape=None,channel=None,figOffset=1,cmap=None,padding=0.01):
    # Receptive Fields Summary
    try:
        W = layer.W
    except:
        W = layer
    wp = W.eval().transpose();
    if len(np.shape(wp)) < 4:       # Fully connected layer, has no shape
        fields = np.reshape(wp,list(wp.shape[0:-1])+fieldShape) 
    else:           # Convolutional layer already has shape
        features, channels, iy, ix = np.shape(wp)
        if channel is not None:
            fields = wp[:,channel,:,:]
        else:
            fields = np.reshape(wp,[features*channels,iy,ix])

    perRow = int(math.floor(math.sqrt(fields.shape[0])))
    perColumn = int(math.ceil(fields.shape[0]/float(perRow)))

    fig = mpl.figure(figOffset); mpl.clf()

    # Using image grid
    from mpl_toolkits.axes_grid1 import ImageGrid
    grid = ImageGrid(fig,111,nrows_ncols=(perRow,perColumn),axes_pad=padding,cbar_mode='single')
    for i in range(0,np.shape(fields)[0]):
        im = grid[i].imshow(fields[i],cmap=cmap); 

    grid.cbar_axes[0].colorbar(im)
    mpl.title('%s Receptive Fields' % layer.name)

    # old way
    # fields2 = np.vstack([fields,np.zeros([perRow*perColumn-fields.shape[0]] + list(fields.shape[1:]))])
    # tiled = []
    # for i in range(0,perColumn*perRow,perColumn):
    #   tiled.append(np.hstack(fields2[i:i+perColumn]))
    # 
    # tiled = np.vstack(tiled)
    # mpl.figure(figOffset); mpl.clf(); mpl.imshow(tiled,cmap=cmap); mpl.title('%s Receptive Fields' % layer.name); mpl.colorbar();
    mpl.figure(figOffset+1); mpl.clf(); mpl.imshow(np.sum(np.abs(fields),0),cmap=cmap); mpl.title('%s Total Absolute Input Dependency' % layer.name); mpl.colorbar()

def plotOutput(layer,feed_dict,fieldShape=None,channel=None,figOffset=1,cmap=None):
    # Output summary
    try:
        W = layer.output
    except:
        W = layer
    wp = W.eval(feed_dict=feed_dict);
    if len(np.shape(wp)) < 4:       # Fully connected layer, has no shape
        temp = np.zeros(np.product(fieldShape)); temp[0:np.shape(wp.ravel())[0]] = wp.ravel()
        fields = np.reshape(temp,[1]+fieldShape)
    else:           # Convolutional layer already has shape
        wp = np.rollaxis(wp,3,0)
        features, channels, iy,ix = np.shape(wp)
        if channel is not None:
            fields = wp[:,channel,:,:]
        else:
            fields = np.reshape(wp,[features*channels,iy,ix])

    perRow = int(math.floor(math.sqrt(fields.shape[0])))
    perColumn = int(math.ceil(fields.shape[0]/float(perRow)))
    fields2 = np.vstack([fields,np.zeros([perRow*perColumn-fields.shape[0]] + list(fields.shape[1:]))])
    tiled = []
    for i in range(0,perColumn*perRow,perColumn):
        tiled.append(np.hstack(fields2[i:i+perColumn]))

    tiled = np.vstack(tiled)
    if figOffset is not None:
        mpl.figure(figOffset); mpl.clf(); 

    mpl.imshow(tiled,cmap=cmap); mpl.title('%s Output' % layer.name); mpl.colorbar();

def setupOutput(self):
        if len(self.input.get_shape()) > 2:
            input = tf.reshape(self.input,[-1,self.inputShape]) # flatten reduced image into a vector
        else:
            input = self.input
        self.output = tf.matmul(input,self.W)

def setupOutput(self):
        if len(self.input.get_shape()) > 2:
            input = tf.reshape(self.input,[-1,self.inputShape]) # flatten reduced image into a vector
        else:
            input = self.input
        self.output = tf.nn.softmax(tf.matmul(input,self.W) + self.b)

def setupOutput(self):
        if len(self.input.get_shape()) > 2:
            input = tf.reshape(self.input,[-1,self.inputShape]) # flatten reduced image into a vector
        else:
            input = self.input
        self.output = tf.nn.relu(tf.matmul(input,self.W) + self.b)