tensorflow::ops::Conv2D::Attrs

#include <nn_ops.h>

Optional attribute setters for Conv2D.

Summary

Public attributes

data_format_ = "NHWC"
StringPiece
dilations_ = {1, 1, 1, 1}
gtl::ArraySlice< int >
use_cudnn_on_gpu_ = true
bool

Public functions

DataFormat(StringPiece x)
TF_MUST_USE_RESULT Attrs
Specify the data format of the input and output data.
Dilations(const gtl::ArraySlice< int > & x)
TF_MUST_USE_RESULT Attrs
1-D tensor of length 4.
UseCudnnOnGpu(bool x)
TF_MUST_USE_RESULT Attrs
Defaults to true.

Public attributes

data_format_

StringPiece tensorflow::ops::Conv2D::Attrs::data_format_ = "NHWC"

dilations_

gtl::ArraySlice< int > tensorflow::ops::Conv2D::Attrs::dilations_ = {1, 1, 1, 1}

use_cudnn_on_gpu_

bool tensorflow::ops::Conv2D::Attrs::use_cudnn_on_gpu_ = true

Public functions

DataFormat

TF_MUST_USE_RESULT Attrs tensorflow::ops::Conv2D::Attrs::DataFormat(
  StringPiece x
)

Specify the data format of the input and output data.

With the default format "NHWC", the data is stored in the order of: [batch, height, width, channels]. Alternatively, the format could be "NCHW", the data storage order of: [batch, channels, height, width].

Defaults to "NHWC"

Dilations

TF_MUST_USE_RESULT Attrs tensorflow::ops::Conv2D::Attrs::Dilations(
  const gtl::ArraySlice< int > & x
)

1-D tensor of length 4.

The dilation factor for each dimension of input. If set to k > 1, there will be k-1 skipped cells between each filter element on that dimension. The dimension order is determined by the value of data_format, see above for details. Dilations in the batch and depth dimensions must be 1.

Defaults to [1, 1, 1, 1]

UseCudnnOnGpu

TF_MUST_USE_RESULT Attrs tensorflow::ops::Conv2D::Attrs::UseCudnnOnGpu(
  bool x
)

Defaults to true.