I am building an Android application that takes decision with the help of a TensorFlow that takes data collected from device sensors.
What I need is the DETAILED steps to create a TensorFlow Light module and to import it into my Android project which is built using Android Studio.
I have already installed Python 3.6.5 (64-bit) on my PC, and installed TensorFlow using the following command:
C:\> pip3 install --upgrade tensorflow
The needed steps should CLEARLY cover the following:
1. How to create a Neural Network with 8 inputs, 6 neurons in hidden layer and 2 outputs.
2. How to feed the Neural Network with data:
a. How to train it (4000 samples, each with 8 real-number parameters)
b. How to test it (500 sample, each with 8 real-number parameters)
3. How to convert the created Neural Network to TensorFlow Light.
4. How to Import that TensorFlow Light into my Android studio project.
As I have no previous experience in both Python and TensorFlow, each step must be explained in detail.