Python shapes.tensor_dim方法代码示例

# 需要导入模块: import shapes [as 别名]
# 或者: from shapes import tensor_dim [as 别名]
def _AddOutputs(self, prev_layer, out_dims, out_func, num_classes):
    """Adds the output layer and loss function.

    Args:
      prev_layer:  Output of last layer of main network.
      out_dims:    Number of output dimensions, 0, 1 or 2.
      out_func:    Output non-linearity. 's' or 'c'=softmax, 'l'=logistic.
      num_classes: Number of outputs/size of last output dimension.
    """
    height_in = shapes.tensor_dim(prev_layer, dim=1)
    logits, outputs = self._AddOutputLayer(prev_layer, out_dims, out_func,
                                           num_classes)
    if self.mode == 'train':
      # Setup loss for training.
      self.loss = self._AddLossFunction(logits, height_in, out_dims, out_func)
      tf.summary.scalar('loss', self.loss)
    elif out_dims == 0:
      # Be sure the labels match the output, even in eval mode.
      self.labels = tf.slice(self.labels, [0, 0], [-1, 1])
      self.labels = tf.reshape(self.labels, [-1])

    logging.info('Final output=%s', outputs)
    logging.info('Labels tensor=%s', self.labels)
    self.output = outputs 

# 需要导入模块: import shapes [as 别名]
# 或者: from shapes import tensor_dim [as 别名]
def _AddOutputs(self, prev_layer, out_dims, out_func, num_classes):
    """Adds the output layer and loss function.

    Args:
      prev_layer:  Output of last layer of main network.
      out_dims:    Number of output dimensions, 0, 1 or 2.
      out_func:    Output non-linearity. 's' or 'c'=softmax, 'l'=logistic.
      num_classes: Number of outputs/size of last output dimension.
    """
    height_in = shapes.tensor_dim(prev_layer, dim=1)
    logits, outputs = self._AddOutputLayer(prev_layer, out_dims, out_func,
                                           num_classes)
    if self.mode == 'train':
      # Setup loss for training.
      self.loss = self._AddLossFunction(logits, height_in, out_dims, out_func)
      tf.scalar_summary('loss', self.loss, name='loss')
    elif out_dims == 0:
      # Be sure the labels match the output, even in eval mode.
      self.labels = tf.slice(self.labels, [0, 0], [-1, 1])
      self.labels = tf.reshape(self.labels, [-1])

    logging.info('Final output=%s', outputs)
    logging.info('Labels tensor=%s', self.labels)
    self.output = outputs 

# 需要导入模块: import shapes [as 别名]
# 或者: from shapes import tensor_dim [as 别名]
def AddFCLayer(self, prev_layer, index):
    """Parse expression and add Fully Connected Layer.

    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'(F)(s|t|r|l|m)({\w+})?(\d+)')
    m = pattern.match(self.model_str, index)
    if m is None:
      return None, None
    fn = self._NonLinearity(m.group(2))
    name = self._GetLayerName(m.group(0), index, m.group(3))
    depth = int(m.group(4))
    input_depth = shapes.tensor_dim(prev_layer, 1) * shapes.tensor_dim(
        prev_layer, 2) * shapes.tensor_dim(prev_layer, 3)
    # The slim fully connected is actually a 1x1 conv, so we have to crush the
    # dimensions on input.
    # Everything except batch goes to depth, and therefore has to be known.
    shaped = tf.reshape(
        prev_layer, [-1, input_depth], name=name + '_reshape_in')
    output = slim.fully_connected(shaped, depth, activation_fn=fn, scope=name)
    # Width and height are collapsed to 1.
    self.reduction_factors[1] = None
    self.reduction_factors[2] = None
    return tf.reshape(
        output, [shapes.tensor_dim(prev_layer, 0), 1, 1, depth],
        name=name + '_reshape_out'), m.end() 

# 需要导入模块: import shapes [as 别名]
# 或者: from shapes import tensor_dim [as 别名]
def _AddLossFunction(self, logits, height_in, out_dims, out_func):
    """Add the appropriate loss function.

    Args:
      logits:  Pre-softmax/logistic fully-connected output shaped to out_dims.
      height_in:  Height of logits before going into the softmax layer.
      out_dims:   Number of output dimensions, 0, 1 or 2.
      out_func:   Output non-linearity. 's' or 'c'=softmax, 'l'=logistic.

    Returns:
      loss: That which is to be minimized.

    Raises:
      ValueError: if logistic is used.
    """
    if out_func == 'c':
      # Transpose batch to the middle.
      ctc_input = tf.transpose(logits, [1, 0, 2])
      # Compute the widths of each batch element from the input widths.
      widths = self.layers.GetLengths(dim=2, factor=height_in)
      cross_entropy = tf.nn.ctc_loss(ctc_input, self.sparse_labels, widths)
    elif out_func == 's':
      if out_dims == 2:
        self.labels = _PadLabels3d(logits, self.labels)
      elif out_dims == 1:
        self.labels = _PadLabels2d(
            shapes.tensor_dim(
                logits, dim=1), self.labels)
      else:
        self.labels = tf.slice(self.labels, [0, 0], [-1, 1])
        self.labels = tf.reshape(self.labels, [-1])
      cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(
          logits=logits, labels=self.labels, name='xent')
    else:
      # TODO(rays) Labels need an extra dimension for logistic, so different
      # padding functions are needed, as well as a different loss function.
      raise ValueError('Logistic not yet supported!')
    return tf.reduce_sum(cross_entropy) 

# 需要导入模块: import shapes [as 别名]
# 或者: from shapes import tensor_dim [as 别名]
def _AddLossFunction(self, logits, height_in, out_dims, out_func):
    """Add the appropriate loss function.

    Args:
      logits:  Pre-softmax/logistic fully-connected output shaped to out_dims.
      height_in:  Height of logits before going into the softmax layer.
      out_dims:   Number of output dimensions, 0, 1 or 2.
      out_func:   Output non-linearity. 's' or 'c'=softmax, 'l'=logistic.

    Returns:
      loss: That which is to be minimized.

    Raises:
      ValueError: if logistic is used.
    """
    if out_func == 'c':
      # Transpose batch to the middle.
      ctc_input = tf.transpose(logits, [1, 0, 2])
      # Compute the widths of each batch element from the input widths.
      widths = self.layers.GetLengths(dim=2, factor=height_in)
      cross_entropy = tf.nn.ctc_loss(ctc_input, self.sparse_labels, widths)
    elif out_func == 's':
      if out_dims == 2:
        self.labels = _PadLabels3d(logits, self.labels)
      elif out_dims == 1:
        self.labels = _PadLabels2d(
            shapes.tensor_dim(
                logits, dim=1), self.labels)
      else:
        self.labels = tf.slice(self.labels, [0, 0], [-1, 1])
        self.labels = tf.reshape(self.labels, [-1])
      cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(
          logits, self.labels, name='xent')
    else:
      # TODO(rays) Labels need an extra dimension for logistic, so different
      # padding functions are needed, as well as a different loss function.
      raise ValueError('Logistic not yet supported!')
    return tf.reduce_sum(cross_entropy)