Welcome to Raspberry Pi 4B repository

1. Preparation

1.1 Place the TensorFlow installation package into Raspberry Pi through WinSCP software.

We put it into /home/pi directory

1.2 Then, input command: ls to check installation package, as shown below.

2. Install tensorflow

2.1 Enter the following instructions to install pip related tools.

sudo apt-get install python3-pip python3-dev

2.2 Enter the following commands in order.

pip3 install h5py-2.10.0-cp37-cp37m-linux_armv7l.whl

pip3 install grpcio-1.32.0-cp37-cp37m-linux_armv7l.whl

pip3 install tensorflow-1.14.0-cp37-none-linux_armv7l.whl

Because I have already installed h5py, running the system again will tell you: “Requirement already satisfied”.  As shown below.

If it is the first installation, a successful English prompt will appear at the bottom to inform you that the installation was successful.

But it is best to read through the installation report. If there are errors, it may be due to some plugins being disconnected due to network reasons.

You can run the installation instructions repeatedly to remedy them

If the remedy fails, you can try to ignore the error and force the installation, with the following instructions:

pip3 install --ignore-installed tensorflow-1.14.0-cp37-none-linux_armv7l.whl

3. Test tensorflow

After completing the installation, enter the following instructions in sequence on the terminal to check if TensorFlow has been successfully installed.

python3

import tensorflow as tf

tf.__version__

If the installed version number appears, it indicates successful installation, and the test results are as follows.

Enter the following command to query the TensorFlow installation path.

tf__ Path__

Initialize a Tensorflow constant: Hello, Tensorflow string, and name it hello as a calculation module.

hello = tf.constant('Hello,Tensorflow')

sess = tf.Session()                      #Start a session

print(sess.run(hello))                   #Using Session Execution Hello Calculation Module