# Tensorflow : Basic

본 포스팅은 모두를 위한 머신러닝/딥러닝 강의(hunkim.github.io/ml)를 참고하였습니다.

1. What is 'Tensorflow'

  • Open source software library for numerical computation using data flow graphs
  • Using Python

2. Data flow graph

  • Nodes in the graph represent mathematical operations
  • Edges represent the multidimensional data arrays (called tensors) communicated between them
  • Node: 수학적 계산, 데이터 입출력, 데이터 읽기 및 저장
  • Edges: 노드들 간 데이터 입출력 관계
    • 동적 사이즈로 구성된 다차원 데이터 배열(Tensor)을 실어 나름
    • That is why we call it 'Tensorflow'

3. Features

  • Abundant expressions through data flow graph
  • Working CPU/GPU mode without code modification
  • Capabilities from Idea test to Service
  • Automatic differential calculation through calculation structure and goal function definition
  • Supporting Python and C++
  • Supporting Various languages by SWIG

4. Installing

You can install Tensorflow on Mac, Windows and Linux.
There are some mechanism by which you install tensorflow.
  • virtualenv
  • native pip
  • Docker
  • installing from source
In fact, Tensorflow official webpage recommends the virtualenv. But it does not matter what it is. You can pick up what you prefer to.
I installed Tensorflow on Mac. Here is the installation steps.

5. Examples

6. Tensorflow Mechanics

  • importing Tensorflow
    • import tensorflow as if
    • To build model(what we called graph)
  • Make Session, the object for executing model

댓글

댓글 쓰기