Input

tensorflow C++ API

tensorflow::Input

Represents a tensor value that can be used as an operand to anOperation.

Publicfunctions

Input

Input(
  const Output & o
)

All of Input’s constructors are implicit.

Input can be implicitly constructed from the following objects :

• Output: This is so that the output of an Operation can be directly used as the input to a op wrapper, which takes Inputs.
• A scalar, or a multi-dimensional tensor specified as a recursive initializer list. This enables directly passing constants as inputs to op wrappers.
• A Tensor object.

Input

 Input(
  const T & v
)

Input

Input(
  const Initializer & init
)

Input

 Input(
  const Tensor & t
)

Input

Input(
  const std::initializer_list<Initializer>& init
)

Input

 Input(
  const string & name,
  int32 i,
  DataType dt
)

Constructor specifying a node name, index and datatype.
This should only be used for specifying a backward edge, needed by control flow.

data_type

DataType data_type() const

index

int32 index() const

node

Node * node() const

node_name

string node_name() const

status

Status status() const

tensor

const Tensor & tensor() const

Inputblock

Source link :https://github.com/EXPNUNI/enuSpaceTensorflow/blob/master/enuSpaceTensorflow/tf_core.cpp

Argument:

• Scope scope : A Scope object (A scope is generated automatically each page. A scope is not connected.)
• DataType dtype: Types of Input Variables. ( DT_INT8, DT_INT16, DT_INT32,DT_INT64, DT_FLOAT, DT_DOUBLE etc….)
• Tensor initvalue: The value of the input variable. It can be written in python syntax. (ex: {1,2} -> shape: [2], input[1] = 1 input[2] = 2)

Output:

• Input input: return Input type of tensor

UsingMethod

※ Input 블럭에 dtype을 DT_INT32로 동일하게 설정하고, Init Value 부분에 각각(위에서 부터) 5와 7을 넣어주고 실행한 뒤 Add블럭의 result_z부분에 결과 값이 나온 화면

※ Input 블럭에 dtype을 DT_INT32로 동일하게 설정하고, Init Value 부분에 각각 python에서 쓰는 shape 형태로(위에서 부터) {5, 3}와 {7, 1}을 넣어주고 실행한 뒤 Add블럭의 result_z부분에 결과 값이 나온 화면