Python tensorflow.cast() Examples

def fprop(self, x, y, **kwargs):
        if self.attack is not None:
            x = x, self.attack(x)
        else:
            x = x,

        # Catching RuntimeError: Variable -= value not supported by tf.eager.
        try:
            y -= self.smoothing * (y - 1. / tf.cast(y.shape[-1], tf.float32))
        except RuntimeError:
            y.assign_sub(self.smoothing * (y - 1. / tf.cast(y.shape[-1],
                                                            tf.float32)))

        logits = [self.model.get_logits(x, **kwargs) for x in x]
        loss = sum(
            softmax_cross_entropy_with_logits(labels=y,
                                              logits=logit)
            for logit in logits)
        warnings.warn("LossCrossEntropy is deprecated, switch to "
                      "CrossEntropy. LossCrossEntropy may be removed on "
                      "or after 2019-03-06.")
        return loss 

def one_hot_encoding(labels, num_classes, scope=None):
  """Transform numeric labels into onehot_labels.

  Args:
    labels: [batch_size] target labels.
    num_classes: total number of classes.
    scope: Optional scope for name_scope.
  Returns:
    one hot encoding of the labels.
  """
  with tf.name_scope(scope, 'OneHotEncoding', [labels]):
    batch_size = labels.get_shape()[0]
    indices = tf.expand_dims(tf.range(0, batch_size), 1)
    labels = tf.cast(tf.expand_dims(labels, 1), indices.dtype)
    concated = tf.concat(axis=1, values=[indices, labels])
    onehot_labels = tf.sparse_to_dense(
        concated, tf.stack([batch_size, num_classes]), 1.0, 0.0)
    onehot_labels.set_shape([batch_size, num_classes])
    return onehot_labels 

def _reshape_instance_masks(self, keys_to_tensors):
    """Reshape instance segmentation masks.

    The instance segmentation masks are reshaped to [num_instances, height,
    width] and cast to boolean type to save memory.

    Args:
      keys_to_tensors: a dictionary from keys to tensors.

    Returns:
      A 3-D boolean tensor of shape [num_instances, height, width].
    """
    masks = keys_to_tensors['image/segmentation/object']
    if isinstance(masks, tf.SparseTensor):
      masks = tf.sparse_tensor_to_dense(masks)
    height = keys_to_tensors['image/height']
    width = keys_to_tensors['image/width']
    to_shape = tf.cast(tf.stack([-1, height, width]), tf.int32)

    return tf.cast(tf.reshape(masks, to_shape), tf.bool) 

def _create_autosummary_var(name, value_expr):
    assert not _autosummary_finalized
    v = tf.cast(value_expr, tf.float32)
    if v.shape.ndims is 0:
        v = [v, np.float32(1.0)]
    elif v.shape.ndims is 1:
        v = [tf.reduce_sum(v), tf.cast(tf.shape(v)[0], tf.float32)]
    else:
        v = [tf.reduce_sum(v), tf.reduce_prod(tf.cast(tf.shape(v), tf.float32))]
    v = tf.cond(tf.is_finite(v[0]), lambda: tf.stack(v), lambda: tf.zeros(2))
    with tf.control_dependencies(None):
        var = tf.Variable(tf.zeros(2)) # [numerator, denominator]
    update_op = tf.cond(tf.is_variable_initialized(var), lambda: tf.assign_add(var, v), lambda: tf.assign(var, v))
    if name in _autosummary_vars:
        _autosummary_vars[name].append(var)
    else:
        _autosummary_vars[name] = [var]
    return update_op

#----------------------------------------------------------------------------
# Call filewriter.add_summary() with all summaries in the default graph,
# automatically finalizing and merging them on the first call. 

def _create_regression_weights(self, match):
    """Set regression weight for each anchor.

    Only positive anchors are set to contribute to the regression loss, so this
    method returns a weight of 1 for every positive anchor and 0 for every
    negative anchor.

    Args:
      match: a matcher.Match object that provides a matching between anchors
        and groundtruth boxes.

    Returns:
      reg_weights: a float32 tensor with shape [num_anchors] representing
        regression weights
    """
    reg_weights = tf.cast(match.matched_column_indicator(), tf.float32)
    return reg_weights 

def scale(keypoints, y_scale, x_scale, scope=None):
  """Scales keypoint coordinates in x and y dimensions.

  Args:
    keypoints: a tensor of shape [num_instances, num_keypoints, 2]
    y_scale: (float) scalar tensor
    x_scale: (float) scalar tensor
    scope: name scope.

  Returns:
    new_keypoints: a tensor of shape [num_instances, num_keypoints, 2]
  """
  with tf.name_scope(scope, 'Scale'):
    y_scale = tf.cast(y_scale, tf.float32)
    x_scale = tf.cast(x_scale, tf.float32)
    new_keypoints = keypoints * [[[y_scale, x_scale]]]
    return new_keypoints 

def minibatch_stddev_layer(x, group_size=4):
    with tf.variable_scope('MinibatchStddev'):
        group_size = tf.minimum(group_size, tf.shape(x)[0])     # Minibatch must be divisible by (or smaller than) group_size.
        s = x.shape                                             # [NCHW]  Input shape.
        y = tf.reshape(x, [group_size, -1, s[1], s[2], s[3]])   # [GMCHW] Split minibatch into M groups of size G.
        y = tf.cast(y, tf.float32)                              # [GMCHW] Cast to FP32.
        y -= tf.reduce_mean(y, axis=0, keep_dims=True)           # [GMCHW] Subtract mean over group.
        y = tf.reduce_mean(tf.square(y), axis=0)                # [MCHW]  Calc variance over group.
        y = tf.sqrt(y + 1e-8)                                   # [MCHW]  Calc stddev over group.
        y = tf.reduce_mean(y, axis=[1,2,3], keep_dims=True)      # [M111]  Take average over fmaps and pixels.
        y = tf.cast(y, x.dtype)                                 # [M111]  Cast back to original data type.
        y = tf.tile(y, [group_size, 1, s[2], s[3]])             # [N1HW]  Replicate over group and pixels.
        return tf.concat([x, y], axis=1)                        # [NCHW]  Append as new fmap.

#----------------------------------------------------------------------------
# Generator network used in the paper. 

def fprop(self, x, y, **kwargs):
        kwargs.update(self.kwargs)
        if self.attack is not None:
            x = x, self.attack(x)
        else:
            x = x,

        # Catching RuntimeError: Variable -= value not supported by tf.eager.
        try:
            y -= self.smoothing * (y - 1. / tf.cast(y.shape[-1], y.dtype))
        except RuntimeError:
            y.assign_sub(self.smoothing * (y - 1. / tf.cast(y.shape[-1],
                                                            y.dtype)))

        logits = [self.model.get_logits(x, **kwargs) for x in x]
        loss = sum(
            tf.reduce_mean(softmax_cross_entropy_with_logits(labels=y,
                                                             logits=logit))
            for logit in logits)
        return loss 

def from_float32_to_uint8(
        tensor,
        tensor_key='tensor',
        min_key='min',
        max_key='max'):
    """

    :param tensor:
    :param tensor_key:
    :param min_key:
    :param max_key:
    :returns:
    """
    tensor_min = tf.reduce_min(tensor)
    tensor_max = tf.reduce_max(tensor)
    return {
        tensor_key: tf.cast(
            (tensor - tensor_min) / (tensor_max - tensor_min + 1e-16)
            * 255.9999, dtype=tf.uint8),
        min_key: tensor_min,
        max_key: tensor_max
    } 

def time_stretch(
        spectrogram,
        factor=1.0,
        method=tf.image.ResizeMethod.BILINEAR):
    """ Time stretch a spectrogram preserving shape in tensorflow. Note that
    this is an approximation in the frequency domain.

    :param spectrogram: Input spectrogram to be time stretched as tensor.
    :param factor: (Optional) Time stretch factor, must be >0, default to 1.
    :param mehtod: (Optional) Interpolation method, default to BILINEAR.
    :returns: Time stretched spectrogram as tensor with same shape.
    """
    T = tf.shape(spectrogram)[0]
    T_ts = tf.cast(tf.cast(T, tf.float32) * factor, tf.int32)[0]
    F = tf.shape(spectrogram)[1]
    ts_spec = tf.image.resize_images(
        spectrogram,
        [T_ts, F],
        method=method,
        align_corners=True)
    return tf.image.resize_image_with_crop_or_pad(ts_spec, T, F) 

def pitch_shift(
        spectrogram,
        semitone_shift=0.0,
        method=tf.image.ResizeMethod.BILINEAR):
    """ Pitch shift a spectrogram preserving shape in tensorflow. Note that
    this is an approximation in the frequency domain.

    :param spectrogram: Input spectrogram to be pitch shifted as tensor.
    :param semitone_shift: (Optional) Pitch shift in semitone, default to 0.0.
    :param mehtod: (Optional) Interpolation method, default to BILINEAR.
    :returns: Pitch shifted spectrogram (same shape as spectrogram).
    """
    factor = 2 ** (semitone_shift / 12.)
    T = tf.shape(spectrogram)[0]
    F = tf.shape(spectrogram)[1]
    F_ps = tf.cast(tf.cast(F, tf.float32) * factor, tf.int32)[0]
    ps_spec = tf.image.resize_images(
        spectrogram,
        [T, F_ps],
        method=method,
        align_corners=True)
    paddings = [[0, 0], [0, tf.maximum(0, F - F_ps)], [0, 0]]
    return tf.pad(ps_spec[:, :F, :], paddings, 'CONSTANT') 

def batch_of_random_bools(batch_size, n):
  """Return a batch of random "boolean" numbers.

  Args:
    batch_size:  Batch size dimension of returned tensor.
    n:  number of entries per batch.

  Returns:
    A [batch_size, n] tensor of "boolean" numbers, where each number is
    preresented as -1 or 1.
  """

  as_int = tf.random_uniform(
      [batch_size, n], minval=0, maxval=2, dtype=tf.int32)
  expanded_range = (as_int * 2) - 1
  return tf.cast(expanded_range, tf.float32) 

def loop_decode(self):
        # decoder_initial_state: Tuple Tensor (c,h) of size [batch_size x cell.state_size]
        # decoder_first_input: Tensor [batch_size x cell.state_size]

        # Loop the decoding process and collect results
        s,i = self.decoder_initial_state,  tf.cast(self.decoder_first_input,tf.float32)
        for step in range(self.seq_length):
            s, i = self.decode(s,i,step)

        # Return to start
        self.positions.append(self.first_city)

        # Stack visited indices
        self.positions=tf.stack(self.positions,axis=1)  # [Batch,seq_length+1]

        # Sum log_softmax over output steps
        self.log_softmax=tf.add_n(self.log_softmax)  # [Batch,seq_length]

        # Stack attending & pointing distribution
        self.attending=tf.stack(self.attending,axis=1) # [Batch,seq_length,seq_length]
        self.pointing=tf.stack(self.pointing,axis=1) # [Batch,seq_length,seq_length]
        
        # Return stacked lists of visited_indices and log_softmax for backprop
        return self.positions,self.log_softmax 

def compute_first_or_last(self, select, first=True):
    #perform first ot last operation on row select with probabilistic row selection
    answer = tf.zeros_like(select)
    running_sum = tf.zeros([self.batch_size, 1], self.data_type)
    for i in range(self.max_elements):
      if (first):
        current = tf.slice(select, [0, i], [self.batch_size, 1])
      else:
        current = tf.slice(select, [0, self.max_elements - 1 - i],
                           [self.batch_size, 1])
      curr_prob = current * (1 - running_sum)
      curr_prob = curr_prob * tf.cast(curr_prob >= 0.0, self.data_type)
      running_sum += curr_prob
      temp_ans = []
      curr_prob = tf.expand_dims(tf.reshape(curr_prob, [self.batch_size]), 0)
      for i_ans in range(self.max_elements):
        if (not (first) and i_ans == self.max_elements - 1 - i):
          temp_ans.append(curr_prob)
        elif (first and i_ans == i):
          temp_ans.append(curr_prob)
        else:
          temp_ans.append(tf.zeros_like(curr_prob))
      temp_ans = tf.transpose(tf.concat(axis=0, values=temp_ans))
      answer += temp_ans
    return answer 

def char_predictions(self, chars_logit):
    """Returns confidence scores (softmax values) for predicted characters.

    Args:
      chars_logit: chars logits, a tensor with shape
        [batch_size x seq_length x num_char_classes]

    Returns:
      A tuple (ids, log_prob, scores), where:
        ids - predicted characters, a int32 tensor with shape
          [batch_size x seq_length];
        log_prob - a log probability of all characters, a float tensor with
          shape [batch_size, seq_length, num_char_classes];
        scores - corresponding confidence scores for characters, a float
        tensor
          with shape [batch_size x seq_length].
    """
    log_prob = utils.logits_to_log_prob(chars_logit)
    ids = tf.to_int32(tf.argmax(log_prob, dimension=2), name='predicted_chars')
    mask = tf.cast(
        slim.one_hot_encoding(ids, self._params.num_char_classes), tf.bool)
    all_scores = tf.nn.softmax(chars_logit)
    selected_scores = tf.boolean_mask(all_scores, mask, name='char_scores')
    scores = tf.reshape(selected_scores, shape=(-1, self._params.seq_length))
    return ids, log_prob, scores 

def loss(logits, labels):
  """Add L2Loss to all the trainable variables.

  Add summary for "Loss" and "Loss/avg".
  Args:
    logits: Logits from inference().
    labels: Labels from distorted_inputs or inputs(). 1-D tensor
            of shape [batch_size]

  Returns:
    Loss tensor of type float.
  """
  # Calculate the average cross entropy loss across the batch.
  labels = tf.cast(labels, tf.int64)
  cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(
      labels=labels, logits=logits, name='cross_entropy_per_example')
  cross_entropy_mean = tf.reduce_mean(cross_entropy, name='cross_entropy')
  tf.add_to_collection('losses', cross_entropy_mean)

  # The total loss is defined as the cross entropy loss plus all of the weight
  # decay terms (L2 loss).
  return tf.add_n(tf.get_collection('losses'), name='total_loss') 

def inputs(eval_data):
  """Construct input for CIFAR evaluation using the Reader ops.

  Args:
    eval_data: bool, indicating if one should use the train or eval data set.

  Returns:
    images: Images. 4D tensor of [batch_size, IMAGE_SIZE, IMAGE_SIZE, 3] size.
    labels: Labels. 1D tensor of [batch_size] size.

  Raises:
    ValueError: If no data_dir
  """
  if not FLAGS.data_dir:
    raise ValueError('Please supply a data_dir')
  data_dir = os.path.join(FLAGS.data_dir, 'cifar-10-batches-bin')
  images, labels = cifar10_input.inputs(eval_data=eval_data,
                                        data_dir=data_dir,
                                        batch_size=FLAGS.batch_size)
  if FLAGS.use_fp16:
    images = tf.cast(images, tf.float16)
    labels = tf.cast(labels, tf.float16)
  return images, labels 

def distorted_inputs():
  """Construct distorted input for CIFAR training using the Reader ops.

  Returns:
    images: Images. 4D tensor of [batch_size, IMAGE_SIZE, IMAGE_SIZE, 3] size.
    labels: Labels. 1D tensor of [batch_size] size.

  Raises:
    ValueError: If no data_dir
  """
  if not FLAGS.data_dir:
    raise ValueError('Please supply a data_dir')
  data_dir = os.path.join(FLAGS.data_dir, 'cifar-10-batches-bin')
  images, labels = cifar10_input.distorted_inputs(data_dir=data_dir,
                                                  batch_size=FLAGS.batch_size)
  if FLAGS.use_fp16:
    images = tf.cast(images, tf.float16)
    labels = tf.cast(labels, tf.float16)
  return images, labels 

def accumulate_privacy_spending(self, eps_delta, unused_sigma,
                                  num_examples):
    """Accumulate the privacy spending.

    Currently only support approximate privacy. Here we assume we use Gaussian
    noise on randomly sampled batch so we get better composition: 1. the per
    batch privacy is computed using privacy amplication via sampling bound;
    2. the composition is done using the composition with Gaussian noise.
    TODO(liqzhang) Add a link to a document that describes the bounds used.

    Args:
      eps_delta: EpsDelta pair which can be tensors.
      unused_sigma: the noise sigma. Unused for this accountant.
      num_examples: the number of examples involved.
    Returns:
      a TensorFlow operation for updating the privacy spending.
    """

    eps, delta = eps_delta
    with tf.control_dependencies(
        [tf.Assert(tf.greater(delta, 0),
                   ["delta needs to be greater than 0"])]):
      amortize_ratio = (tf.cast(num_examples, tf.float32) * 1.0 /
                        self._total_examples)
      # Use privacy amplification via sampling bound.
      # See Lemma 2.2 in http://arxiv.org/pdf/1405.7085v2.pdf
      # TODO(liqzhang) Add a link to a document with formal statement
      # and proof.
      amortize_eps = tf.reshape(tf.log(1.0 + amortize_ratio * (
          tf.exp(eps) - 1.0)), [1])
      amortize_delta = tf.reshape(amortize_ratio * delta, [1])
      return tf.group(*[tf.assign_add(self._eps_squared_sum,
                                      tf.square(amortize_eps)),
                        tf.assign_add(self._delta_sum, amortize_delta)]) 

def AddReShape(self, prev_layer, index):
    """Reshapes the input tensor by moving each (x_scale,y_scale) rectangle to.

       the depth dimension. NOTE that the TF convention is that inputs are
       [batch, y, x, depth].

    Args:
      prev_layer: Input tensor.
      index:      Position in model_str to start parsing

    Returns:
      Output tensor, end index in model_str.
    """
    pattern = re.compile(R'(S)(?:{(\w)})?(\d+)\((\d+)x(\d+)\)(\d+),(\d+)')
    m = pattern.match(self.model_str, index)
    if m is None:
      return None, None
    name = self._GetLayerName(m.group(0), index, m.group(2))
    src_dim = int(m.group(3))
    part_a = int(m.group(4))
    part_b = int(m.group(5))
    dest_dim_a = int(m.group(6))
    dest_dim_b = int(m.group(7))
    if part_a == 0:
      part_a = -1
    if part_b == 0:
      part_b = -1
    prev_shape = tf.shape(prev_layer)
    layer = shapes.transposing_reshape(
        prev_layer, src_dim, part_a, part_b, dest_dim_a, dest_dim_b, name=name)
    # Compute scale factors.
    result_shape = tf.shape(layer)
    for i in xrange(len(self.reduction_factors)):
      if self.reduction_factors[i] is not None:
        factor1 = tf.cast(self.reduction_factors[i], tf.float32)
        factor2 = tf.cast(prev_shape[i], tf.float32)
        divisor = tf.cast(result_shape[i], tf.float32)
        self.reduction_factors[i] = tf.div(tf.multiply(factor1, factor2), divisor)
    return layer, m.end() 

def compute_lookup_error(self, val):
    #computes lookup error.
    cond = tf.equal(self.batch_print_answer, val)
    inter = tf.where(
        cond, self.init_print_error,
        tf.tile(
            tf.reshape(tf.constant(1e10, self.data_type), [1, 1, 1]), [
                self.batch_size, self.utility.FLAGS.max_word_cols +
                self.utility.FLAGS.max_number_cols,
                self.utility.FLAGS.max_elements
            ]))
    return tf.reduce_min(tf.reduce_min(inter, 1), 1) * tf.cast(
        tf.greater(
            tf.reduce_sum(tf.reduce_sum(tf.cast(cond, self.data_type), 1), 1),
            0.0), self.data_type) 

def compute_max_or_min(self, select, maxi=True):
    #computes the argmax and argmin of a column with probabilistic row selection
    answer = tf.zeros([
        self.batch_size, self.num_cols + self.num_word_cols, self.max_elements
    ], self.data_type)
    sum_prob = tf.zeros([self.batch_size, self.num_cols + self.num_word_cols],
                        self.data_type)
    for j in range(self.max_elements):
      if (maxi):
        curr_pos = j
      else:
        curr_pos = self.max_elements - 1 - j
      select_index = tf.slice(self.full_processed_sorted_index_column,
                              [0, 0, curr_pos], [self.batch_size, -1, 1])
      select_mask = tf.equal(
          tf.tile(
              tf.expand_dims(
                  tf.tile(
                      tf.expand_dims(tf.range(self.max_elements), 0),
                      [self.batch_size, 1]), 1),
              [1, self.num_cols + self.num_word_cols, 1]), select_index)
      curr_prob = tf.expand_dims(select, 1) * tf.cast(
          select_mask, self.data_type) * self.select_bad_number_mask
      curr_prob = curr_prob * tf.expand_dims((1 - sum_prob), 2)
      curr_prob = curr_prob * tf.expand_dims(
          tf.cast((1 - sum_prob) > 0.0, self.data_type), 2)
      answer = tf.where(select_mask, curr_prob, answer)
      sum_prob += tf.reduce_sum(curr_prob, 2)
    return answer 

def make_hard_softmax(self, softmax):
    #converts soft selection to hard selection. used at test time
    cond = tf.equal(
        softmax, tf.reshape(tf.reduce_max(softmax, 1), [self.batch_size, 1]))
    softmax = tf.where(
        cond, tf.fill(tf.shape(softmax), 1.0), tf.fill(tf.shape(softmax), 0.0))
    softmax = tf.cast(softmax, self.data_type)
    return softmax 

def padded_one_hot_encoding(indices, depth, left_pad):
  """Returns a zero padded one-hot tensor.

  This function converts a sparse representation of indices (e.g., [4]) to a
  zero padded one-hot representation (e.g., [0, 0, 0, 0, 1] with depth = 4 and
  left_pad = 1). If `indices` is empty, the result will simply be a tensor of
  shape (0, depth + left_pad). If depth = 0, then this function just returns
  `None`.

  Args:
    indices: an integer tensor of shape [num_indices].
    depth: depth for the one-hot tensor (integer).
    left_pad: number of zeros to left pad the one-hot tensor with (integer).

  Returns:
    padded_onehot: a tensor with shape (num_indices, depth + left_pad). Returns
      `None` if the depth is zero.

  Raises:
    ValueError: if `indices` does not have rank 1 or if `left_pad` or `depth are
      either negative or non-integers.

  TODO: add runtime checks for depth and indices.
  """
  if depth < 0 or not isinstance(depth, (int, long) if six.PY2 else int):
    raise ValueError('`depth` must be a non-negative integer.')
  if left_pad < 0 or not isinstance(left_pad, (int, long) if six.PY2 else int):
    raise ValueError('`left_pad` must be a non-negative integer.')
  if depth == 0:
    return None
  if len(indices.get_shape().as_list()) != 1:
    raise ValueError('`indices` must have rank 1')

  def one_hot_and_pad():
    one_hot = tf.cast(tf.one_hot(tf.cast(indices, tf.int64), depth,
                                 on_value=1, off_value=0), tf.float32)
    return tf.pad(one_hot, [[0, 0], [left_pad, 0]], mode='CONSTANT')
  result = tf.cond(tf.greater(tf.size(indices), 0), one_hot_and_pad,
                   lambda: tf.zeros((depth + left_pad, 0)))
  return tf.reshape(result, [-1, depth + left_pad]) 

def sequence_accuracy(predictions, targets, rej_char, streaming=False):
  """Computes sequence level accuracy.

  Both input tensors should have the same shape: [batch_size x seq_length].

  Args:
    predictions: predicted character classes.
    targets: ground truth character classes.
    rej_char: the character id used to mark empty element (end of sequence).
    streaming: if True, uses the streaming mean from the slim.metric module.

  Returns:
    a update_ops for execution and value tensor whose value on evaluation
    returns the total sequence accuracy.
  """

  with tf.variable_scope('SequenceAccuracy'):
    predictions.get_shape().assert_is_compatible_with(targets.get_shape())

    targets = tf.to_int32(targets)
    const_rej_char = tf.constant(
        rej_char, shape=targets.get_shape(), dtype=tf.int32)
    include_mask = tf.not_equal(targets, const_rej_char)
    include_predictions = tf.to_int32(
        tf.where(include_mask, predictions,
                 tf.zeros_like(predictions) + rej_char))
    correct_chars = tf.to_float(tf.equal(include_predictions, targets))
    correct_chars_counts = tf.cast(
        tf.reduce_sum(correct_chars, reduction_indices=[1]), dtype=tf.int32)
    target_length = targets.get_shape().dims[1].value
    target_chars_counts = tf.constant(
        target_length, shape=correct_chars_counts.get_shape())
    accuracy_per_example = tf.to_float(
        tf.equal(correct_chars_counts, target_chars_counts))
    if streaming:
      return tf.contrib.metrics.streaming_mean(accuracy_per_example)
    else:
      return tf.reduce_mean(accuracy_per_example) 

def _sample_box_classifier_minibatch(self,
                                       proposal_boxlist,
                                       groundtruth_boxlist,
                                       groundtruth_classes_with_background):
    """Samples a mini-batch of proposals to be sent to the box classifier.

    Helper function for self._postprocess_rpn.

    Args:
      proposal_boxlist: A BoxList containing K proposal boxes in absolute
        coordinates.
      groundtruth_boxlist: A Boxlist containing N groundtruth object boxes in
        absolute coordinates.
      groundtruth_classes_with_background: A tensor with shape
        `[N, self.num_classes + 1]` representing groundtruth classes. The
        classes are assumed to be k-hot encoded, and include background as the
        zero-th class.

    Returns:
      a BoxList contained sampled proposals.
    """
    (cls_targets, cls_weights, _, _, _) = self._detector_target_assigner.assign(
        proposal_boxlist, groundtruth_boxlist,
        groundtruth_classes_with_background)
    # Selects all boxes as candidates if none of them is selected according
    # to cls_weights. This could happen as boxes within certain IOU ranges
    # are ignored. If triggered, the selected boxes will still be ignored
    # during loss computation.
    cls_weights += tf.to_float(tf.equal(tf.reduce_sum(cls_weights), 0))
    positive_indicator = tf.greater(tf.argmax(cls_targets, axis=1), 0)
    sampled_indices = self._second_stage_sampler.subsample(
        tf.cast(cls_weights, tf.bool),
        self._second_stage_batch_size,
        positive_indicator)
    return box_list_ops.boolean_mask(proposal_boxlist, sampled_indices) 

def compute_mfcc(audio, **kwargs):
    """
    Compute the MFCC for a given audio waveform. This is
    identical to how DeepSpeech does it, but does it all in
    TensorFlow so that we can differentiate through it.
    """

    batch_size, size = audio.get_shape().as_list()
    audio = tf.cast(audio, tf.float32)

    # 1. Pre-emphasizer, a high-pass filter
    audio = tf.concat((audio[:, :1], audio[:, 1:] - 0.97*audio[:, :-1], np.zeros((batch_size,1000),dtype=np.float32)), 1)

    # 2. windowing into frames of 320 samples, overlapping
    windowed = tf.stack([audio[:, i:i+400] for i in range(0,size-320,160)],1)

    # 3. Take the FFT to convert to frequency space
    ffted = tf.spectral.rfft(windowed, [512])
    ffted = 1.0 / 512 * tf.square(tf.abs(ffted))

    # 4. Compute the Mel windowing of the FFT
    energy = tf.reduce_sum(ffted,axis=2)+1e-30
    filters = np.load("filterbanks.npy").T
    feat = tf.matmul(ffted, np.array([filters]*batch_size,dtype=np.float32))+1e-30

    # 5. Take the DCT again, because why not
    feat = tf.log(feat)
    feat = tf.spectral.dct(feat, type=2, norm='ortho')[:,:,:26]

    # 6. Amplify high frequencies for some reason
    _,nframes,ncoeff = feat.get_shape().as_list()
    n = np.arange(ncoeff)
    lift = 1 + (22/2.)*np.sin(np.pi*n/22)
    feat = lift*feat
    width = feat.get_shape().as_list()[1]

    # 7. And now stick the energy next to the features
    feat = tf.concat((tf.reshape(tf.log(energy),(-1,width,1)), feat[:, :, 1:]), axis=2)
    
    return feat 

def parser(self, value):
    """Parse a Cifar10 record from value.

    Output images are in [height, width, depth] layout.
    """
    # Dimensions of the images in the CIFAR-10 dataset.
    # See http://www.cs.toronto.edu/~kriz/cifar.html for a description of the
    # input format.
    label_bytes = 1
    image_bytes = HEIGHT * WIDTH * DEPTH
    # Every record consists of a label followed by the image, with a
    # fixed number of bytes for each.
    record_bytes = label_bytes + image_bytes

    # Convert from a string to a vector of uint8 that is record_bytes long.
    record_as_bytes = tf.decode_raw(value, tf.uint8)

    # The first bytes represent the label, which we convert from
    # uint8->int32.
    label = tf.cast(
        tf.strided_slice(record_as_bytes, [0], [label_bytes]), tf.int32)

    label.set_shape([1])

    # The remaining bytes after the label represent the image, which
    # we reshape from [depth * height * width] to [depth, height, width].
    depth_major = tf.reshape(
        tf.strided_slice(record_as_bytes, [label_bytes], [record_bytes]),
        [3, 32, 32])
    # Convert from [depth, height, width] to [height, width, depth].
    # This puts data in a compatible layout with TF image preprocessing APIs.
    image = tf.transpose(depth_major, [1, 2, 0])

    # Do custom preprocessing here.
    image = self.preprocess(image)

    return image, label 

def optimize(self, loss):
    """Build the graph to optimize the loss function."""

    # Optimizer nodes.
    # Linear learning rate decay.
    opts = self._options
    words_to_train = float(opts.words_per_epoch * opts.epochs_to_train)
    lr = opts.learning_rate * tf.maximum(
        0.0001, 1.0 - tf.cast(self._words, tf.float32) / words_to_train)
    self._lr = lr
    optimizer = tf.train.GradientDescentOptimizer(lr)
    train = optimizer.minimize(loss,
                               global_step=self.global_step,
                               gate_gradients=optimizer.GATE_NONE)
    self._train = train 

def inverse_sigmoid_decay(k, global_step_op):
  with tf.name_scope('inverse_sigmoid_decay'):
    k = tf.constant(k, dtype=tf.float32)
    tmp = k*tf.exp(-tf.cast(global_step_op, tf.float32)/k)
    tmp = tmp / (1. + tmp)
  return tmp