Shulou(Shulou.com)06/01 Report--

Today, I would like to talk to you about why there is TensorFlow, many people may not know much about it. In order to make you understand better, the editor summarized the following content for you. I hope you can get something according to this article.

01

-

Why is there a TensorFlow?

Although the mathematical concepts behind deep learning have been around for decades, the program class libraries used to create and train these deep models have only emerged in recent years.

Unfortunately, most class libraries make tradeoffs between flexibility and actual production value. Easy-to-use class libraries are extremely valuable for scientific research, because this makes it easier for scientists to develop new model frameworks, but if they take into account the actual production, they are often too slow to train and have no practical application value. On the other hand, fast and efficient class libraries are feasible to deploy to distributed hardware, but they are only for specific neural networks and are not suitable for newly invented networks.

This leaves decision makers wondering: do we still have to do scientific research with inflexible class libraries if we don't repeat the same code? Or do scientific research and production each use its own different set of libraries?

If we choose the former, we may not be able to test different kinds of neural networks, if we choose the latter, we have to maintain two different sets of code, can we have enough energy to do this?

TensorFlow is born to solve the mission of how the emerging neural network architecture can really be used in a production environment.

02

-

What is TensorFlow?

TensorFlow is an open source software library that uses data flow diagrams to do numerical calculations.

03

-

Why choose TensorFlow?

The reasons are as follows:

Has been very popular.

Clearer API

Can be actually used in the production environment

Features of TensorFlow:

Python API

Easy deployment: one or more CPUs or GPUs, server, client, mobile phone

Flexibility: use Android, windows,iOS,Linux

Visualization: TensorBoard can see the data flow diagram

Checkpoints: easy to manage experiments

Automatic differentiation

The community is huge

Many good projects have used or are using TensorFlow

04

-

TensorFlow-based project

Projects based on TensorFlow:

DeepMind's WaveNet Text to speech

Google Brain's Magenta project that uses machine learning to create compelling art and music

Neural Style Translation

Major Improvement to Google Translate

The following figure is Neural style Translation:

Generative handwriting (LSTM mixed density network combined with TensorFlow):

WaveNet Text to speech (generation model for original audio)

After reading the above, do you have any further understanding of why there is TensorFlow? If you want to know more knowledge or related content, please follow the industry information channel, thank you for your support.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.