From AlphaGo victory over Lee Shishi, Google artificial intelligence has attracted much attention. Artificial intelligence seems to be far from us, depth learning algorithm looks very complicated.But in fact, look at your phone on the voice assistant, the camera’s face recognition, today’s headlines to help you screen out the news automatically, as well as the music software “Daily Recommended” and so on. All kinds of AI has already entered all aspects of our lives. It can be said that this is an era of AI.
In fact, as early as the end of last year, Google has open-source the depth learning system TensorFlow which is used for making AlphaGo, I believe many people are very interested in this powerful machine learning system.
To install TensorFlow on Firefly-RK3399. If you want to play with artificial intelligence, you can follow the tutorial to try.
1. Install Ubuntu System
Flash the ubuntu system image to Firefly-RK3399. I upgraded the Ubuntu single system: Firefly-RK3399_xubuntu1604_201707041527.img, you can go to the official website to download it:
[u]http://en.t-firefly.com/en/firenow/Firefly_RK3399/download/[/u]
Note, if your Firefly-RK3399 is the default dual system firmware, you can switch directly to the Ubuntu system and skip this step:
[u]http://wiki.t-firefly.com/index.php/Firefly-RK3399/Starter_guide/en[/u]
update successfully.
2. Install pip3
After entering the system, use ALT + CTRL + T to bring up the terminal and enter
sudo apt install python3-pip
3. Use git to download tensorflow-1.1.0rc1-cp35-cp35m-linux_aarch64.whl
Install git
sudo apt-get install git-core
Download tensorflow-1.1.0rc1-cp35-cp35m-linux_aarch64.whl
git clone https://github.com/JohnnyChan8/TensorFlow4FireflyRK3399.git
4. Use pip3 to install Tensorflow
cd TensorFlow4FireflyRK3399
pip3 install tensorflow-1.1.0rc1-cp35-cp35m-linux_aarch64.whl
Install successfully
5. Test Tensorflow
After completing the above steps, TensorFlow has been installed to Firefly-RK3399, and now through a simple routine to test the program input
python3
import tensorflow as tf
hello = tf.constant('Hello world!')
sess = tf.Session()
print(sess.run(hello))
This time will print
b’Hello world!’
Summary
TensorFlow as a machine learning tool has been open source for a year, with Google’s influence TensorFlow has become a popular open source machine learning, in just a year or so, in GitHub has produced a large number of cases based on TensorFlow, such as images Recognition, natural language processing, small AlphaGo, machine writing music, and so interesting items.
Artificial intelligence is really not far away from us, open source TensorFlow let us have the opportunity to create our own artificial intelligence.