Python tensorflow.python.ops.math_ops.to_int32() Examples

def loss(self, data, labels):
    """The loss to minimize while training."""

    if self.is_regression:
      diff = self.training_inference_graph(data) - math_ops.to_float(labels)
      mean_squared_error = math_ops.reduce_mean(diff * diff)
      root_mean_squared_error = math_ops.sqrt(mean_squared_error, name="loss")
      loss = root_mean_squared_error
    else:
      loss = math_ops.reduce_mean(
          nn_ops.sparse_softmax_cross_entropy_with_logits(
              labels=array_ops.squeeze(math_ops.to_int32(labels)),
              logits=self.training_inference_graph(data)),
          name="loss")
    if self.regularizer:
      loss += layers.apply_regularization(self.regularizer,
                                          variables.trainable_variables())
    return loss 

def loss(self, data, labels):
    """The loss to minimize while training."""

    if self.is_regression:
      diff = self.training_inference_graph(data) - math_ops.to_float(labels)
      mean_squared_error = math_ops.reduce_mean(diff * diff)
      root_mean_squared_error = math_ops.sqrt(mean_squared_error, name="loss")
      loss = root_mean_squared_error
    else:
      loss = math_ops.reduce_mean(
          nn_ops.sparse_softmax_cross_entropy_with_logits(
              self.training_inference_graph(data),
              array_ops.squeeze(math_ops.to_int32(labels))),
          name="loss")
    if self.regularizer:
      loss += layers.apply_regularization(self.regularizer,
                                          variables.trainable_variables())
    return loss 

def average_impurity(self):
    """Constructs a TF graph for evaluating the average leaf impurity of a tree.

    If in regression mode, this is the leaf variance. If in classification mode,
    this is the gini impurity.

    Returns:
      The last op in the graph.
    """
    children = array_ops.squeeze(array_ops.slice(
        self.variables.tree, [0, 0], [-1, 1]), squeeze_dims=[1])
    is_leaf = math_ops.equal(constants.LEAF_NODE, children)
    leaves = math_ops.to_int32(array_ops.squeeze(array_ops.where(is_leaf),
                                                 squeeze_dims=[1]))
    counts = array_ops.gather(self.variables.node_sums, leaves)
    gini = self._weighted_gini(counts)
    # Guard against step 1, when there often are no leaves yet.
    def impurity():
      return gini
    # Since average impurity can be used for loss, when there's no data just
    # return a big number so that loss always decreases.
    def big():
      return array_ops.ones_like(gini, dtype=dtypes.float32) * 10000000.
    return control_flow_ops.cond(math_ops.greater(
        array_ops.shape(leaves)[0], 0), impurity, big) 

def average_impurity(self):
    """Constructs a TF graph for evaluating the average leaf impurity of a tree.

    If in regression mode, this is the leaf variance. If in classification mode,
    this is the gini impurity.

    Returns:
      The last op in the graph.
    """
    children = array_ops.squeeze(array_ops.slice(
        self.variables.tree, [0, 0], [-1, 1]), squeeze_dims=[1])
    is_leaf = math_ops.equal(constants.LEAF_NODE, children)
    leaves = math_ops.to_int32(array_ops.squeeze(array_ops.where(is_leaf),
                                                 squeeze_dims=[1]))
    counts = array_ops.gather(self.variables.node_sums, leaves)
    gini = self._weighted_gini(counts)
    # Guard against step 1, when there often are no leaves yet.
    def impurity():
      return gini
    # Since average impurity can be used for loss, when there's no data just
    # return a big number so that loss always decreases.
    def big():
      return array_ops.ones_like(gini, dtype=dtypes.float32) * 10000000.
    return control_flow_ops.cond(math_ops.greater(
        array_ops.shape(leaves)[0], 0), impurity, big) 

def average_impurity(self):
    """Constructs a TF graph for evaluating the average leaf impurity of a tree.

    If in regression mode, this is the leaf variance. If in classification mode,
    this is the gini impurity.

    Returns:
      The last op in the graph.
    """
    children = array_ops.squeeze(array_ops.slice(
        self.variables.tree, [0, 0], [-1, 1]), squeeze_dims=[1])
    is_leaf = math_ops.equal(constants.LEAF_NODE, children)
    leaves = math_ops.to_int32(array_ops.squeeze(array_ops.where(is_leaf),
                                                 squeeze_dims=[1]))
    counts = array_ops.gather(self.variables.node_sums, leaves)
    gini = self._weighted_gini(counts)
    # Guard against step 1, when there often are no leaves yet.
    def impurity():
      return gini
    # Since average impurity can be used for loss, when there's no data just
    # return a big number so that loss always decreases.
    def big():
      return array_ops.ones_like(gini, dtype=dtypes.float32) * 10000000.
    return control_flow_ops.cond(math_ops.greater(
        array_ops.shape(leaves)[0], 0), impurity, big) 

def _GatherGrad(op, grad):
  """Gradient for Gather op."""
  # params can be large, so colocate the shape calculation with it.
  #
  # params can be very large for sparse model, array_ops.shape raises
  # exception on the Windows platform when any dimension is larger than
  # int32. params_shape is not used in optimizer apply_sparse gradients,
  # so it's fine to convert it back to int32 regardless of truncation.
  params = op.inputs[0]
  with ops.colocate_with(params):
    params_shape = array_ops.shape(params, out_type=ops.dtypes.int64)
    params_shape = math_ops.to_int32(params_shape)

  # Build appropriately shaped IndexedSlices
  indices = op.inputs[1]
  size = array_ops.expand_dims(array_ops.size(indices), 0)
  values_shape = array_ops.concat([size, params_shape[1:]], 0)
  values = array_ops.reshape(grad, values_shape)
  indices = array_ops.reshape(indices, size)
  return [ops.IndexedSlices(values, indices, params_shape), None] 

def loss(self, data, labels):
    """The loss to minimize while training."""

    if self.is_regression:
      diff = self.training_inference_graph(data) - math_ops.to_float(labels)
      mean_squared_error = math_ops.reduce_mean(diff * diff)
      root_mean_squared_error = math_ops.sqrt(mean_squared_error, name="loss")
      loss = root_mean_squared_error
    else:
      loss = math_ops.reduce_mean(
          nn_ops.sparse_softmax_cross_entropy_with_logits(
              labels=array_ops.squeeze(math_ops.to_int32(labels)),
              logits=self.training_inference_graph(data)),
          name="loss")
    if self.regularizer:
      loss += layers.apply_regularization(self.regularizer,
                                          variables.trainable_variables())
    return loss 

def average_impurity(self):
    """Constructs a TF graph for evaluating the average leaf impurity of a tree.

    If in regression mode, this is the leaf variance. If in classification mode,
    this is the gini impurity.

    Returns:
      The last op in the graph.
    """
    children = array_ops.squeeze(array_ops.slice(
        self.variables.tree, [0, 0], [-1, 1]), squeeze_dims=[1])
    is_leaf = math_ops.equal(constants.LEAF_NODE, children)
    leaves = math_ops.to_int32(array_ops.squeeze(array_ops.where(is_leaf),
                                                 squeeze_dims=[1]))
    counts = array_ops.gather(self.variables.node_sums, leaves)
    gini = self._weighted_gini(counts)
    # Guard against step 1, when there often are no leaves yet.
    def impurity():
      return gini
    # Since average impurity can be used for loss, when there's no data just
    # return a big number so that loss always decreases.
    def big():
      return array_ops.ones_like(gini, dtype=dtypes.float32) * 10000000.
    return control_flow_ops.cond(math_ops.greater(
        array_ops.shape(leaves)[0], 0), impurity, big) 

def average_impurity(self):
    """Constructs a TF graph for evaluating the average leaf impurity of a tree.

    If in regression mode, this is the leaf variance. If in classification mode,
    this is the gini impurity.

    Returns:
      The last op in the graph.
    """
    children = array_ops.squeeze(array_ops.slice(
        self.variables.tree, [0, 0], [-1, 1]), squeeze_dims=[1])
    is_leaf = math_ops.equal(constants.LEAF_NODE, children)
    leaves = math_ops.to_int32(array_ops.squeeze(array_ops.where(is_leaf),
                                                 squeeze_dims=[1]))
    counts = array_ops.gather(self.variables.node_sums, leaves)
    gini = self._weighted_gini(counts)
    # Guard against step 1, when there often are no leaves yet.
    def impurity():
      return gini
    # Since average impurity can be used for loss, when there's no data just
    # return a big number so that loss always decreases.
    def big():
      return array_ops.ones_like(gini, dtype=dtypes.float32) * 10000000.
    return control_flow_ops.cond(math_ops.greater(
        array_ops.shape(leaves)[0], 0), impurity, big) 

def loss(self, data, labels):
    """The loss to minimize while training."""

    if self.is_regression:
      diff = self.training_inference_graph(data) - math_ops.to_float(labels)
      mean_squared_error = math_ops.reduce_mean(diff * diff)
      root_mean_squared_error = math_ops.sqrt(mean_squared_error, name="loss")
      loss = root_mean_squared_error
    else:
      loss = math_ops.reduce_mean(
          nn_ops.sparse_softmax_cross_entropy_with_logits(
              labels=array_ops.squeeze(math_ops.to_int32(labels)),
              logits=self.training_inference_graph(data)),
          name="loss")
    if self.regularizer:
      loss += layers.apply_regularization(self.regularizer,
                                          variables.trainable_variables())
    return loss 

def _select_which_to_enqueue(tensor_list, keep_input):
  """Select which examples to enqueue based on vector `keep_input`."""
  select_i = math_ops.to_int32(keep_input)
  tensor_list = [
      data_flow_ops.dynamic_partition(x, select_i, num_partitions=2)[1]
      for x in tensor_list]
  return tensor_list 

def _top_k_generator(k):
  def _top_k(probabilities, targets):
    targets = math_ops.to_int32(targets)
    if targets.get_shape().ndims > 1:
      targets = array_ops.squeeze(targets, squeeze_dims=[1])
    return metric_ops.streaming_mean(nn.in_top_k(probabilities, targets, k))
  return _top_k 

def one_hot_wrapper(num_classes, loss_fn):
  """Some loss functions take one-hot labels."""
  def _loss(probs, targets):
    if targets.get_shape().ndims > 1:
      targets = array_ops.squeeze(targets, squeeze_dims=[1])
    one_hot_labels = array_ops.one_hot(
        math_ops.to_int32(targets),
        num_classes,
        on_value=1.,
        off_value=0.,
        dtype=dtypes.float32)
    return loss_fn(probs, one_hot_labels)
  return _loss 

def _compute_zeroone_score(labels, predictions):
  zeroone_score = math_ops.to_float(
      math_ops.equal(
          math_ops.reduce_sum(
              math_ops.to_int32(math_ops.equal(labels, predictions))),
          array_ops.shape(labels)[0]))
  return zeroone_score 

def _softmax_entropy(probabilities, targets, weights=None):
  return metric_ops.streaming_mean(
      losses.sparse_softmax_cross_entropy(probabilities,
                                          math_ops.to_int32(targets)),
      weights=weights) 

def ctc_batch_cost(y_true, y_pred, input_length, label_length):
  """Runs CTC loss algorithm on each batch element.

  Arguments:
      y_true: tensor `(samples, max_string_length)`
          containing the truth labels.
      y_pred: tensor `(samples, time_steps, num_categories)`
          containing the prediction, or output of the softmax.
      input_length: tensor `(samples, 1)` containing the sequence length for
          each batch item in `y_pred`.
      label_length: tensor `(samples, 1)` containing the sequence length for
          each batch item in `y_true`.

  Returns:
      Tensor with shape (samples,1) containing the
          CTC loss of each element.
  """
  label_length = math_ops.to_int32(array_ops.squeeze(label_length))
  input_length = math_ops.to_int32(array_ops.squeeze(input_length))
  sparse_labels = math_ops.to_int32(
      ctc_label_dense_to_sparse(y_true, label_length))

  y_pred = math_ops.log(array_ops.transpose(y_pred, perm=[1, 0, 2]) + 1e-8)

  return array_ops.expand_dims(
      ctc.ctc_loss(
          inputs=y_pred, labels=sparse_labels, sequence_length=input_length), 1) 

def _softmax_entropy(probabilities, targets, weights=None):
  return metric_ops.streaming_mean(losses.sparse_softmax_cross_entropy(
      probabilities, math_ops.to_int32(targets)),
                                   weights=weights) 

def _squeeze_and_onehot(targets, depth):
  targets = array_ops.squeeze(targets, squeeze_dims=[1])
  return array_ops.one_hot(math_ops.to_int32(targets), depth) 

def one_hot_wrapper(num_classes, loss_fn):
  """Some loss functions take one-hot labels."""
  def _loss(probs, targets):
    one_hot_labels = array_ops.one_hot(
        math_ops.to_int32(targets), num_classes,
        on_value=1., off_value=0., dtype=dtypes.float32)
    return loss_fn(probs, one_hot_labels)
  return _loss 

def get_best(self, n):
    """Return the indices and values of the n highest scores in the TopN."""

    def refresh_shortlist():
      """Update the shortlist with the highest scores in id_to_score."""
      new_scores, new_ids = nn_ops.top_k(self.id_to_score, self.shortlist_size)
      smallest_new_score = math_ops.reduce_min(new_scores)
      new_length = math_ops.reduce_sum(
          math_ops.to_int32(math_ops.greater(new_scores, dtypes.float32.min)))
      u1 = self.sl_ids.assign(
          math_ops.to_int64(array_ops.concat([[new_length], new_ids], 0)))
      u2 = self.sl_scores.assign(
          array_ops.concat([[smallest_new_score], new_scores], 0))
      self.last_ops = [u1, u2]
      return control_flow_ops.group(u1, u2)

    # We only need to refresh the shortlist if n is greater than the
    # current shortlist size (which is stored in sl_ids[0]).
    with ops.control_dependencies(self.last_ops):
      cond_op = control_flow_ops.cond(n > self.sl_ids[0], refresh_shortlist,
                                      control_flow_ops.no_op)
      with ops.control_dependencies([cond_op]):
        topk_values, topk_indices = nn_ops.top_k(
            self.sl_scores,
            math_ops.minimum(n, math_ops.to_int32(self.sl_ids[0])))
        # topk_indices are the indices into the shortlist, we want to return
        # the indices into id_to_score
        gathered_indices = array_ops.gather(self.sl_ids, topk_indices)
        return gathered_indices, topk_values 

def one_hot_mask(labels, num_classes, scope=None):
  """Compute 1-hot encodings for masks.

  Given a label image, this computes the one hot encoding at
  each pixel.

  Args:
    labels: (batch_size, width, height, 1) tensor containing labels.
    num_classes: number of classes
    scope: optional scope name

  Returns:
    Tensor of shape (batch_size, width, height, num_classes) with
    a 1-hot encoding.
  """
  with ops.name_scope(scope, "OneHotMask", [labels]):
    height, width, depth = _shape(labels)
    assert depth == 1
    sparse_labels = math_ops.to_int32(array_ops.reshape(labels, [-1, 1]))
    sparse_size, _ = _shape(sparse_labels)
    indices = array_ops.reshape(math_ops.range(0, sparse_size, 1), [-1, 1])
    concated = array_ops.concat([indices, sparse_labels], 1)
    dense_result = sparse_ops.sparse_to_dense(concated,
                                              [sparse_size, num_classes], 1.0,
                                              0.0)
    result = array_ops.reshape(dense_result, [height, width, num_classes])
    return result 

def one_hot_mask(labels, num_classes, scope=None):
  """Compute 1-hot encodings for masks.

  Given a label image, this computes the one hot encoding at
  each pixel.

  Args:
    labels: (batch_size, width, height, 1) tensor containing labels.
    num_classes: number of classes
    scope: optional scope name

  Returns:
    Tensor of shape (batch_size, width, height, num_classes) with
    a 1-hot encoding.
  """
  with ops.name_scope(scope, "OneHotMask", [labels]):
    height, width, depth = _shape(labels)
    assert depth == 1
    sparse_labels = math_ops.to_int32(array_ops.reshape(labels, [-1, 1]))
    sparse_size, _ = _shape(sparse_labels)
    indices = array_ops.reshape(math_ops.range(0, sparse_size, 1), [-1, 1])
    concated = array_ops.concat([indices, sparse_labels], 1)
    dense_result = sparse_ops.sparse_to_dense(concated,
                                              [sparse_size, num_classes], 1.0,
                                              0.0)
    result = array_ops.reshape(dense_result, [height, width, num_classes])
    return result 

def _softmax_entropy(probabilities, targets, weights=None):
  return metric_ops.streaming_mean(
      losses.sparse_softmax_cross_entropy(probabilities,
                                          math_ops.to_int32(targets)),
      weights=weights) 

def _top_k_generator(k):
  def _top_k(probabilities, targets):
    targets = math_ops.to_int32(targets)
    if targets.get_shape().ndims > 1:
      targets = array_ops.squeeze(targets, squeeze_dims=[1])
    return metric_ops.streaming_mean(nn.in_top_k(probabilities, targets, k))
  return _top_k 

def get_best(self, n):
    """Return the indices and values of the n highest scores in the TopN."""

    def refresh_shortlist():
      """Update the shortlist with the highest scores in id_to_score."""
      new_scores, new_ids = nn_ops.top_k(self.id_to_score, self.shortlist_size)
      smallest_new_score = math_ops.reduce_min(new_scores)
      new_length = math_ops.reduce_sum(
          math_ops.to_int32(math_ops.greater(new_scores, dtypes.float32.min)))
      u1 = self.sl_ids.assign(
          math_ops.to_int64(array_ops.concat([[new_length], new_ids], 0)))
      u2 = self.sl_scores.assign(
          array_ops.concat([[smallest_new_score], new_scores], 0))
      self.last_ops = [u1, u2]
      return control_flow_ops.group(u1, u2)

    # We only need to refresh the shortlist if n is greater than the
    # current shortlist size (which is stored in sl_ids[0]).
    with ops.control_dependencies(self.last_ops):
      cond_op = control_flow_ops.cond(n > self.sl_ids[0], refresh_shortlist,
                                      control_flow_ops.no_op)
      with ops.control_dependencies([cond_op]):
        topk_values, topk_indices = nn_ops.top_k(
            self.sl_scores,
            math_ops.minimum(n, math_ops.to_int32(self.sl_ids[0])))
        # topk_indices are the indices into the shortlist, we want to return
        # the indices into id_to_score
        gathered_indices = array_ops.gather(self.sl_ids, topk_indices)
        return gathered_indices, topk_values 

def one_hot_wrapper(num_classes, loss_fn):
  """Some loss functions take one-hot labels."""
  def _loss(probs, targets):
    if targets.get_shape().ndims > 1:
      targets = array_ops.squeeze(targets, squeeze_dims=[1])
    one_hot_labels = array_ops.one_hot(
        math_ops.to_int32(targets),
        num_classes,
        on_value=1.,
        off_value=0.,
        dtype=dtypes.float32)
    return loss_fn(probs, one_hot_labels)
  return _loss 

def _class_labels_streaming_mean(labels, weights, class_id):
  return metrics_lib.streaming_mean(
      array_ops.where(
          math_ops.equal(
              math_ops.to_int32(class_id), math_ops.to_int32(labels)),
          array_ops.ones_like(labels), array_ops.zeros_like(labels)),
      weights=weights) 

def _class_predictions_streaming_mean(predictions, weights, class_id):
  return metrics_lib.streaming_mean(
      array_ops.where(
          math_ops.equal(
              math_ops.to_int32(class_id), math_ops.to_int32(predictions)),
          array_ops.ones_like(predictions),
          array_ops.zeros_like(predictions)),
      weights=weights) 

def one_hot_mask(labels, num_classes, scope=None):
  """Compute 1-hot encodings for masks.

  Given a label image, this computes the one hot encoding at
  each pixel.

  Args:
    labels: (batch_size, width, height, 1) tensor containing labels.
    num_classes: number of classes
    scope: optional scope name

  Returns:
    Tensor of shape (batch_size, width, height, num_classes) with
    a 1-hot encoding.
  """
  with ops.name_scope(scope, "OneHotMask", [labels]):
    height, width, depth = _shape(labels)
    assert depth == 1
    sparse_labels = math_ops.to_int32(array_ops.reshape(labels, [-1, 1]))
    sparse_size, _ = _shape(sparse_labels)
    indices = array_ops.reshape(math_ops.range(0, sparse_size, 1), [-1, 1])
    concated = array_ops.concat([indices, sparse_labels], 1)
    dense_result = sparse_ops.sparse_to_dense(concated,
                                              [sparse_size, num_classes], 1.0,
                                              0.0)
    result = array_ops.reshape(dense_result, [height, width, num_classes])
    return result 

def _softmax_entropy(probabilities, targets, weights=None):
  return metric_ops.streaming_mean(
      losses.sparse_softmax_cross_entropy(probabilities,
                                          math_ops.to_int32(targets)),
      weights=weights)