This application has no documentation, yet. The goal is to create some kind of editor for CNNs.

The software requires Python because the deep learing part is implemented in Pytorch. While you can use a virtual environment for Python I found it easier to have a local Python installation in the Project that can be easily deployed. The local Python installation should be located at Python/PythonInstallation

Recommended procedure to install Python:

First install conda here

Open Miniconda (Windows) or a new terminal (Linux). Move to the Python/PythonInstallation folder using the cd command. Then type conda create -p . python=3.8 Follow the instructions and proceed until Python is installed. Important: The python executable has to be located directly in Python/PythonInstallation Upgrade pip. ./python.exe -m pip install --upgrade pip Then, install the required packages. If you desire GPU support, visit https://pytorch.org/ to get the correct extra-index-url for your system and install cuda https://developer.nvidia.com/cuda-downloads. ./python.exe -m pip install torch torchvision torchgeometry opencv-python websockets matplotlib scikit-learn --extra-index-url https://download.pytorch.org/whl/cu116 The new Python installation has to know where the Python scripts for this application are located. In the Lib/site-packages directory of the local Python installation add a new text file "mypath.pth". Write a single line into the file that contains the path to the Scripts folder. For example: PathToYourProjectFolder/CNNWorkshopVR/Python/Scripts Then you can start the Godot application.