Python v1.boolean_mask方法代码示例

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def test_actnorm(self):
    _, x_mask, _ = self.get_data()
    x = tf.random_normal((BATCH_SIZE, TARGET_LENGTH, N_CHANNELS),
                         mean=50.0, stddev=10.0, dtype=DTYPE)
    x_act, logabsdet = glow.actnorm(
        "actnorm", x, x_mask, inverse=False, init=True)

    x_act_nopad = tf.boolean_mask(x_act, x_mask)
    x_mean, x_var = tf.nn.moments(x_act_nopad, axes=[0])
    self.evaluate(tf.global_variables_initializer())
    x, x_act, logabsdet, x_mean, x_var = (
        self.evaluate([x, x_act, logabsdet, x_mean, x_var]))
    self.assertEqual(x_act.shape, (BATCH_SIZE, TARGET_LENGTH, N_CHANNELS))
    self.assertEqual(logabsdet.shape, (BATCH_SIZE,))
    self.assertTrue(np.allclose(x_mean, 0.0, atol=1e-5))
    self.assertTrue(np.allclose(x_var, 1.0, atol=1e-5)) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def __init__(self, c, d=None, prune_irrelevant=True, collapse=True):
    """Builds a linear specification module."""
    super(LinearSpecification, self).__init__(name='specs', collapse=collapse)
    # c has shape [batch_size, num_specifications, num_outputs]
    # d has shape [batch_size, num_specifications]
    # Some specifications may be irrelevant (not a function of the output).
    # We automatically remove them for clarity. We expect the number of
    # irrelevant specs to be equal for all elements of a batch.
    # Shape is [batch_size, num_specifications]
    if prune_irrelevant:
      irrelevant = tf.equal(tf.reduce_sum(
          tf.cast(tf.abs(c) > 1e-6, tf.int32), axis=-1, keepdims=True), 0)
      batch_size = tf.shape(c)[0]
      num_outputs = tf.shape(c)[2]
      irrelevant = tf.tile(irrelevant, [1, 1, num_outputs])
      self._c = tf.reshape(
          tf.boolean_mask(c, tf.logical_not(irrelevant)),
          [batch_size, -1, num_outputs])
    else:
      self._c = c
    self._d = d 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def sparse_reduce(sp_tensor, rank, agg_fn="sum", axis=-1):
  """Reduce SparseTensor along the given axis.

  Args:
    sp_tensor: SparseTensor of arbitrary rank.
    rank: Integer rank of the sparse tensor.
    agg_fn: Reduce function for aggregation.
    axis: Integer specifying axis to sum over.

  Returns:
    sp_tensor: SparseTensor of one less rank.
  """
  if axis < 0:
    axis = rank + axis
  axes_to_keep = tf.one_hot(
      axis, rank, on_value=False, off_value=True, dtype=tf.bool)
  indices_to_keep = tf.boolean_mask(sp_tensor.indices, axes_to_keep, axis=1)
  new_shape = tf.boolean_mask(sp_tensor.dense_shape, axes_to_keep)
  indices_to_keep.set_shape([None, rank - 1])
  indices, values = aggregate_sparse_indices(
      indices_to_keep, sp_tensor.values, new_shape, agg_fn=agg_fn)
  return tf.sparse.reorder(tf.SparseTensor(indices, values, new_shape)) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def noise_span_to_sentinel(tokens, noise_mask, vocabulary):
  """Replace each run of consecutive noise tokens with a single sentinel.

  Args:
    tokens: a 1d integer Tensor
    noise_mask: a boolean Tensor with the same shape as tokens
    vocabulary: a vocabulary.Vocabulary
  Returns:
    a Tensor with the same shape and dtype as tokens
  """
  tokens = tf.where_v2(noise_mask,
                       tf.cast(sentinel_id(vocabulary), tokens.dtype),
                       tokens)
  prev_token_is_noise = tf.pad(noise_mask[:-1], [[1, 0]])
  subsequent_noise_tokens = tf.logical_and(noise_mask, prev_token_is_noise)
  return tf.boolean_mask(tokens, tf.logical_not(subsequent_noise_tokens)) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def permute_noise_tokens(tokens, noise_mask, unused_vocabulary):
  """Permute the noise tokens, keeping the non-noise tokens where they are.

  Args:
    tokens: a 1d integer Tensor
    noise_mask: a boolean Tensor with the same shape as tokens
    unused_vocabulary: a vocabulary.Vocabulary
  Returns:
    a Tensor with the same shape and dtype as tokens
  """
  masked_only = tf.boolean_mask(tokens, noise_mask)
  permuted = tf.random.shuffle(masked_only)
  # pad to avoid errors when it has size 0
  permuted = tf.pad(permuted, [[0, 1]])
  indices = tf.cumsum(tf.cast(noise_mask, tf.int32), exclusive=True)
  return tf.where_v2(noise_mask,
                     tf.gather(permuted, indices),
                     tokens) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def _get_refined_encodings_for_postitive_class(
      self, refined_box_encodings, flat_cls_targets_with_background,
      batch_size):
    # We only predict refined location encodings for the non background
    # classes, but we now pad it to make it compatible with the class
    # predictions
    refined_box_encodings_with_background = tf.pad(refined_box_encodings,
                                                   [[0, 0], [1, 0], [0, 0]])
    refined_box_encodings_masked_by_class_targets = (
        box_list_ops.boolean_mask(
            box_list.BoxList(
                tf.reshape(refined_box_encodings_with_background,
                           [-1, self._box_coder.code_size])),
            tf.reshape(tf.greater(flat_cls_targets_with_background, 0), [-1]),
            use_static_shapes=self._use_static_shapes,
            indicator_sum=batch_size * self.max_num_proposals
            if self._use_static_shapes else None).get())
    return tf.reshape(
        refined_box_encodings_masked_by_class_targets, [
            batch_size, self.max_num_proposals,
            self._box_coder.code_size
        ]) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def additive_coupling(
    name, x, x_mask, inverse, split_dim, identity_first, init,
    decoder_self_attention_bias=None, **kwargs):
  """Additive coupling transform layer."""
  hparams = kwargs["hparams"]
  batch_size, length, n_channels = common_layers.shape_list(x)
  assert hparams.scale_width > 0.0 and hparams.scale_width < 1.0
  with tf.variable_scope(name, reuse=tf.AUTO_REUSE):
    x_id, x_tr, _, n_transform, bias, mask = gops.split_coupling(
        x, x_mask, split_dim, identity_first, decoder_self_attention_bias)
    z_id = x_id

    loc = gops.transformer_decoder_block(
        "theta_tr",
        n_layers=hparams.n_layers_transform_params,
        x=x_id,
        x_mask=mask,
        output_size=n_transform,
        init=init,
        decoder_self_attention_bias=bias,
        **kwargs)
    if not inverse:
      z_tr = x_tr + loc
    else:
      z_tr = x_tr - loc
    logabsdet = tf.constant(0.0, dtype=tf.float32)

    tf.summary.histogram("_loc", tf.boolean_mask(loc, mask))
    result = gops.join_coupling(z_id, z_tr, split_dim, identity_first)
    result = tf.reshape(result, [batch_size, length, n_channels])
    return result, logabsdet 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def test_dense_weightnorm(self):
    x, x_mask = self.get_data()
    x = tf.random_normal((BATCH_SIZE, TARGET_LENGTH, HIDDEN_SIZE),
                         mean=0.0, stddev=1.0)
    y = gops.dense_weightnorm("wn", x, N_CHANNELS, x_mask,
                              init_scale=1.0, init=True)

    y_nopad = tf.boolean_mask(y, x_mask)
    mean, var = tf.nn.moments(y_nopad, axes=[0])
    self.evaluate(tf.global_variables_initializer())
    x, x_mask, y, y_nopad, mean, var = (
        self.evaluate([x, x_mask, y, y_nopad, mean, var]))
    self.assertEqual(y.shape, (BATCH_SIZE, TARGET_LENGTH, N_CHANNELS))
    self.assertTrue(np.allclose(mean, 0.0, atol=1e-5))
    self.assertTrue(np.allclose(var, 1.0, atol=1e-5)) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def test_empty_x_results_in_empty_output(self):
    # Empty x is common if someone masks their input with tf.boolean_mask in
    # order to drop missing entries, and in a particular batch all entries are
    # missing.
    with self.cached_session():
      x = np.array([]).reshape(0, 3)
      self.assertEqual(0, array_ops.size(x).eval())
      y = _layers.legacy_fully_connected(x, 2, activation_fn=nn_ops.softmax)
      variables_lib.global_variables_initializer().run()
      expected_y = np.array([]).reshape(0, 2)
      np.testing.assert_array_equal(expected_y, y.eval()) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def select_slate_optimal(slate_size, s_no_click, s, q):
  """Selects the slate using exhaustive search.

  This algorithm corresponds to the method "OS" in
  Ie et al. https://arxiv.org/abs/1905.12767.

  Args:
    slate_size: int, the size of the recommendation slate.
    s_no_click: float tensor, the score for not clicking any document.
    s: [num_of_documents] tensor, the scores for clicking documents.
    q: [num_of_documents] tensor, the predicted q values for documents.

  Returns:
    [slate_size] tensor, the selected slate.
  """

  num_candidates = s.shape.as_list()[0]

  # Obtain all possible slates given current docs in the candidate set.
  mesh_args = [list(range(num_candidates))] * slate_size
  slates = tf.stack(tf.meshgrid(*mesh_args), axis=-1)
  slates = tf.reshape(slates, shape=(-1, slate_size))

  # Filter slates that include duplicates to ensure each document is picked
  # at most once.
  unique_mask = tf.map_fn(
      lambda x: tf.equal(tf.size(input=x), tf.size(input=tf.unique(x)[0])),
      slates,
      dtype=tf.bool)
  slates = tf.boolean_mask(tensor=slates, mask=unique_mask)

  slate_q_values = tf.gather(s * q, slates)
  slate_scores = tf.gather(s, slates)
  slate_normalizer = tf.reduce_sum(
      input_tensor=slate_scores, axis=1) + s_no_click

  slate_q_values = slate_q_values / tf.expand_dims(slate_normalizer, 1)
  slate_sum_q_values = tf.reduce_sum(input_tensor=slate_q_values, axis=1)
  max_q_slate_index = tf.argmax(input=slate_sum_q_values)
  return tf.gather(slates, max_q_slate_index, axis=0) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def reduce_concat_tokens(dataset,
                         feature_key='targets',
                         batch_size=128,
                         **unused_kwargs):
  """Token-preprocessor to concatenate multiple unrelated documents.

  If we want to generate examples of exactly the right length,
  (to avoid wasting space on padding), then we use this function, folowed by
  split_tokens.

  Args:
    dataset: a tf.data.Dataset with dictionaries containing the key feature_key.
    feature_key: an string
    batch_size: an integer - how many documents to concatenate into one

  Returns:
    a dataset
  """
  dataset = dataset.map(lambda x: {feature_key: x[feature_key]},
                        num_parallel_calls=tf.data.experimental.AUTOTUNE)
  dataset = dataset.padded_batch(batch_size, padded_shapes={feature_key: [-1]})
  def _my_fn(x):
    tokens = tf.reshape(x[feature_key], [-1])
    # strip padding
    tokens = tf.boolean_mask(tokens, tf.cast(tokens, tf.bool))
    return {feature_key: tokens}
  return dataset.map(_my_fn, num_parallel_calls=tf.data.experimental.AUTOTUNE) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def noise_span_to_unique_sentinel(tokens, noise_mask, vocabulary):
  """Replace each run of consecutive noise tokens with a different sentinel.

  The idea here is to be able to align the dropped spans in the inputs
  with the markers in the targets.

  We want to generate training examples like
  "We hold X to be Y that" -> "X these truths Y self evident Z"

  Sentinels assigned in decreasing order within the sequence starting at
  vocabulary.size - 1.  That is, we appropriate the last tokens in the
  vocabulary for additional use as sentinels.

  TODO(noam): we may want to try enlarging the vocabulary and leaving room
  for the sentinels instead.  However, this requires enlarging the embedding
  tables in the model, so that is a bigger change.

  Args:
    tokens: a 1d integer Tensor
    noise_mask: a boolean Tensor with the same shape as tokens
    vocabulary: a vocabulary.Vocabulary
  Returns:
    a Tensor with the same shape and dtype as tokens
  """
  vocab_size = vocabulary.vocab_size
  prev_token_is_noise = tf.pad(noise_mask[:-1], [[1, 0]])

  first_noise_tokens = tf.logical_and(
      noise_mask, tf.logical_not(prev_token_is_noise))
  subsequent_noise_tokens = tf.logical_and(noise_mask, prev_token_is_noise)

  sentinel = vocab_size - tf.cumsum(tf.cast(first_noise_tokens, tokens.dtype))

  tokens = tf.where_v2(first_noise_tokens, sentinel, tokens)
  return tf.boolean_mask(tokens, tf.logical_not(subsequent_noise_tokens)) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def drop_noise_tokens(tokens, noise_mask, unused_vocabulary):
  """Drop noise tokens without inserting a sentinel.

  Args:
    tokens: a 1d integer Tensor
    noise_mask: a boolean Tensor with the same shape as tokens
    unused_vocabulary: a vocabulary.Vocabulary
  Returns:
    a Tensor with the same shape and dtype as tokens
  """
  return tf.boolean_mask(tokens, tf.logical_not(noise_mask)) 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def trim_zeros_graph(boxes, name='trim_zeros'):
    """Often boxes are represented with matrices of shape [N, 4] and
    are padded with zeros. This removes zero boxes.

    boxes: [N, 4] matrix of boxes.
    non_zeros: [N] a 1D boolean mask identifying the rows to keep
    """
    non_zeros = tf.cast(tf.reduce_sum(tf.abs(boxes), axis=1), tf.bool)
    boxes = tf.boolean_mask(boxes, non_zeros, name=name)
    return boxes, non_zeros 

# 需要导入模块: from tensorflow.compat import v1 [as 别名]
# 或者: from tensorflow.compat.v1 import boolean_mask [as 别名]
def _slice_with_actions(embeddings, actions):
  """Slice a Tensor.

  Take embeddings of the form [batch_size, num_actions, embed_dim]
  and actions of the form [batch_size, 1], and return the sliced embeddings
  like embeddings[:, actions, :].

  Args:
    embeddings: Tensor of embeddings to index.
    actions: int Tensor to use as index into embeddings

  Returns:
    Tensor of embeddings indexed by actions
  """
  batch_size, num_actions = embeddings.get_shape()[:2]

  # Values are the 'values' in a sparse tensor we will be setting
  act_indx = tf.cast(actions, tf.int64)[:, None]
  values = tf.reshape(tf.cast(tf.ones(tf.shape(actions)), tf.bool), [-1])

  # Create a range for each index into the batch
  act_range = tf.range(0, batch_size, dtype=tf.int64)[:, None]
  # Combine this into coordinates with the action indices
  indices = tf.concat([act_range, act_indx], 1)

  actions_mask = tf.SparseTensor(indices, values, [batch_size, num_actions])
  actions_mask = tf.stop_gradient(
      tf.sparse_tensor_to_dense(actions_mask, default_value=False))
  sliced_emb = tf.boolean_mask(embeddings, actions_mask)
  return sliced_emb