If you want to see how the training was done: https://github.com/Coofy/sight-bot-train
Miniconda:
conda create -n perception-tf1.15 pip python=3.7 pyqt=5
pip install tensorflow==1.15
pip install pyautogui
pip install pynput
git clone --depth 1 https://github.com/tensorflow/models
cd models/research/
protoc object_detection/protos/*.proto --python_out=.
pip install .
pip install opencv-python
python main.py
# Force CPU
CUDA_VISIBLE_DEVICES=-1 python main.py