Creating an AI (Windows)
This will walk you through the steps of preparing your system to build an AI model.
Installing Python
First, install Python. Ideally, you should match the Python version being used on your ARDI server.
On Windows systems, we often suggest using the previous major version, as it can take some time for the required libraries to be made available for the Windows platform.
https://www.python.org/downloads/
For simplicity, we normally suggest installing to an easy-to-remember folder, such as C:\Python311 for Python 3.11.
Installing Libraries
Next, you need to install the required libraries.
1) Open a command prompt window (open your Start menu and type 'command prompt' to find it).
2) Type the following command…
pip install tensorflow numpy matplotlib
If the command above fails, you might not have added the Python folder to your path. You can try typing c:\<pythonfolder>\scripts\pip install tensorflow numpy matplotlib instead.
Build your AI
Next, unzip the file you downloaded from the ARDI server into an easy-to-remember folder (for example, 'C:\AIFiles'.
In your command-prompt window, change to that directory. For example…
cd c:\AIFiles
Run the 'build.py' command…
python build.py
As before, if you didn't add Python to your PATH, you might need to enter the full path to your Python executable.
Test your AI
You can test your AI the same way, with the command…
python test.py