What is TensorFlow?

What is TensorFlow?

What is TensorFlow?

TensorFlow is an end-to-end open source platform for learning the machine. It has a broad and flexible ecosystem of tools, libraries and community resources that allows researchers to promote the state of the art in ML and for developers to easily create and implement ML applications.

TensorFlow Architecture

The Tensorflow architecture works in three parts:

  • Pre-processing of the data
  • Builds the model
  • Train and estimate the model

It is called TensorFlow because it receives input as a multidimensional matrix, also known as tensors. You can build a type of flowchart of operations (called Graph) that you want to run on that entry. The entrance enters at one end and then flows through that system of multiple operations and comes out at the other end as an exit. That’s why it’s called TensorFlow because the tensor enters it through a list of operations, and then it comes out the other side.

Tensorflow 1.x

  • Programs in 0.x may not necessarily work in 1.x.
  • Numpy was more similar
  • More stable for array-based computing
  • Two APIs for Java and Go were introduced.
  • The Tensorflow debugger was introduced.
  • new Android demo for the stylization of images based on the camera and the detection of objects.

Computational graphs

A computational graph is a graph of data flow that the TensorFlow represents internally during the execution of an operation. This addressed graph has a set of nodes, each representing an operation and a set of directed arcs, each one representing the data in which the operations are executed. It has two types of edges, one called Normal, which are only structure carriers. of data between the nodes and the second is called Special, which does not load any value, but shows the controlled dependency between two nodes.

Computational Graph

In this graph, we can see two arcs that must be computed and then an arc where the computed value will be assigned, and also between them is the node where the operation is represented and connecting all the knots and arcs are the normal edges. The need for a computational graph is that the TensorFlow programs the most efficient way when evaluating complex expressions in the learning phase that distributes the computational load.

Tensor Data Structure

Tensors are used as basic data structures in the TensorFlow language. The tensors represent the connecting edges in any flowchart called the Data Flow Chart. Tensioners are defined as a multidimensional matrix or list. The sensors are identified by the following three parameters:


The unit of dimensionality described within the tensor is called a range. Identify the number of dimensions of the tensor. A classification of a tensor can be described as the order or n dimensions of a defined tensor.


The number of rows and columns together define the shape of the tensor.


The type describes the type of data assigned to the elements of the Tensor. A user must consider the following activities to build a Tensor –

  • Construct an n-dimensional matrix
  • Convert the n-dimensional matrix.

Applications of TensorFlow

Voice/Sound Recognition

One of the best-known uses of TensorFlow is sound-based applications. With the right data source, neural networks are able to understand audio signals. These are:

• Voice recognition – used mainly in IoT, Automotive, Security and UX / UI

• Voice search – mainly in telecommunications, device manufacturers

• Sentiment analysis – used mainly in CRM

• Fault detection (noise, engine) – used mainly in Automobile and Aviation

Text-Based Applications

Other popular uses of TensorFlow are text-based applications, such as sentimental analysis (CRM, Social Media), Threat Detection (Social Media, Government) and Fraud Detection (Insurance, Finance). Language detection is one of the most popular uses of text-based applications.

We all know Google Translator, which supports more than 100 languages translating from one to another. Evolutionary versions can be used in many cases, such as the translation of legalistic jargon into contracts in simple language.

Image Recognition

Mainly used by manufacturers of social media, telecommunications and cell phones; Image recognition aims to recognize and identify people and objects in images, as well as to understand the content of the image, the detection of movement, the vision of the machine and the grouping of photos, they can also be used in the automotive industries, of aviation and health. and the context.