tensorflow::ops::LRN

Local Response Normalization.

Summary

The 4-D input tensor is treated as a 3-D array of 1-D vectors (along the last dimension), and each vector is normalized independently. Within a given vector, each component is divided by the weighted, squared sum of inputs within depth_radius. In detail,

sqr_sum[a, b, c, d] =
    sum(input[a, b, c, d - depth_radius : d + depth_radius + 1] ** 2)
output = input / (bias + alpha * sqr_sum) ** beta

For details, see Krizhevsky et al., ImageNet classification with deep convolutional neural networks (NIPS 2012).

Args:

• scope: A Scope object
• input: 4-D.

Optional attributes (see Attrs):

• depth_radius: 0-D. Half-width of the 1-D normalization window.
• bias: An offset (usually positive to avoid dividing by 0).
• alpha: A scale factor, usually positive.
• beta: An exponent.

Returns:

• Output: The output tensor.

Public attributes

Operation operation

::tensorflow::Output output

Public functions

 LRN(
  const ::tensorflow::Scope & scope,
  ::tensorflow::Input input
)

 LRN(
  const ::tensorflow::Scope & scope,
  ::tensorflow::Input input,
  const LRN::Attrs & attrs
)

::tensorflow::Node * node() const 

 operator::tensorflow::Input() const 

 operator::tensorflow::Output() const 

Public static functions

Attrs Alpha(
  float x
)

Attrs Beta(
  float x
)

Attrs Bias(
  float x
)

Attrs DepthRadius(
  int64 x
)