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

This article shows you what things to pay attention to in TensorFlow, the content is concise and easy to understand, it will definitely brighten your eyes. I hope you can get something through the detailed introduction of this article.

# 1 this is a powerful machine learning framework

TensorFlow is a machine learning framework, and if you have a lot of data and / or you have mastered the most advanced AI: deep learning, then it may be your best friend. Neural network. Big. It's not a data science Swiss Army knife, it's an industrial lathe. This means that if all you want to do is make a regression through a 20 x 2 spreadsheet, you may stop reading.

But if you're after something big, be happy. TensorFlow has been used to find new planets, prevent blindness by helping doctors screen for diabetic retinopathy, and help save forests by alerting authorities to signs of illegal deforestation. This is the foundation of AlphaGo and Google Cloud Vision, and it is also your best choice. TensorFlow is open source, and you can download it for free and start using it immediately.

The discovery of the planet Kepler-90i with the help of TensorFlow makes the Kepler-90 system the only known system with eight planets in the orbit of a star. No systems with more than eight planets have been found, so I think this means that we are first tied with Kepler-90 (now).

# 2 Strange methods are optional

TensorFlow Eager gives me peace of mind.

If you've tried TensorFlow in the past and because it forces you to run around like academics / aliens instead of developers, come on!

TensorFlow is eager to execute and let you interact with it like a pure Python programmer: all the real-time writing and debugging, rather than holding your breath while building those huge graphics. I'm a recovering scholar myself (and probably an alien), but I've been in love with TF's eagerness to execute since it appeared. So eager to please!

# 3 you can build a neural network line by line

Keras + TensorFlow = easier to build neural networks!

Keras is all about user-friendliness and simple prototyping, and the old TensorFlow is desperate for more. If you like object-oriented thinking, and you like to build neural networks one layer at a time, you will definitely like tf.keras. In the following lines of code, we create a sequential neural network where standard ringtones and whistles are like dropping out of school (reminding me that lyric metaphors about me sometimes drop out, involving staplers and flu).

Oh, you like puzzles, don't you? Patience. Don't think too much about staplers.

# 4 this is not just Python

Well, you've been complaining about TensorFlow's Python monomania for a while. Good news! TensorFlow is not just for Pythonistas. It now runs in multiple languages, from R to Swift to JavaScript.

# 5 you can perform all operations in the browser

Speaking of JavaScript, you can use TensorFlow.js to train and execute the model in the browser. Get out of the nerd in the cool presentation, and I'll still be here when you get back.

Real-time human posture estimation is carried out by using TensorFlow.js in the browser. Turn on the demo on the camera here. Or don't leave your chair. It's up to you.

# 6 Lite version is available for micro devices

Got a clumsy desktop from the museum? toaster? (the same thing? TensorFlow Lite brings model execution to a variety of devices, including mobile and the Internet of things, and reasoning is more than three times faster than the original TensorFlow. Yes, now you can learn about machines on Raspberry Pi or mobile phones. In his speech, Lawrence did a brave thing by demonstrating image classification in real time on an Android simulator in front of thousands of people. And it works.

1.6 seconds? Check! The probability of bananas is over 97%? Check! Toilet paper? Well, I've been to some countries, and I think paper like Lawrence is very useful.

# 7 Professional hardware is getting better

If you are tired of waiting for CPU to complete data training to train your neural network, you can now use hardware designed specifically for Cloud TPU work. T stands for tensor. Just like TensorFlow. Coincidence? I don't think so! A few weeks ago, Google released the alpha version of 3 TPU.

# 8 the new data pipeline has been greatly improved

What are you doing numpy over there? If you want to do this in TensorFlow and then rage-quit, the tf.data namespace now makes your input processing in TensorFlow more expressive and efficient. Tf.data provides you with a fast, flexible and easy-to-use data pipeline that synchronizes with training.

# 9 you don't have to start from scratch

Do you know what is an interesting way to start machine learning? A blank new page in the editor with no mileage sample code. With TensorFlow Hub, you can participate in a more efficient tradition of helping yourself use other people's code and call it your own code (also known as professional software engineering).

TensorFlow Hub is a reusable repository of pre-trained machine learning model components packaged for single-line reuse. Help yourself!

What are the above contents that TensorFlow should pay attention to? have you learned any knowledge or skills? If you want to learn more skills or enrich your knowledge reserve, you are welcome to follow the industry information channel.

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.