Link Search Menu Expand Document

Input::Initializer


tensorflowC++API

tensorflow::Input::Initializer

R Initializer enables constructing an Input object from various kinds of C++ constants such as simple primitive constants and nested initializer lists representing a multi-dimensional array.


Summary

Initializer constructors are all templates, so the aforementioned kinds of C++ constants can be used to construct an Initializer. Initializer stores the value it got constructed with in a Tensor object.


Input::Initializer block

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

Argument:

  • Tensor t: 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::Initializer output: Returns Input :: Initializer. Input :: Initializer is usually used to initialize the placeholder.

UsingMethod

※input_initializer에서 tensor를 직접 생성하여 placeholder를 사용하는 경우

※tensor 블록을 이용하여 placeholder를 사용하는 경우