Plays Set using a strategically-angled laptop webcam to see the board. Cards are classified using a ConvNet, written in PyTorch.
Download pre-trained model and place in a models/ directory at the root of this repository.
Install the following dependencies:
numpy==1.16.0
opencv-python==3.4.3.18
torch==0.4.1
torchvision==0.2.1
scikit-image==0.15.0
pyobjc==5.2
pyttsx3==2.7
python main.py -play to play Set against the bot.
python main.py -check to use the bot to check for sets on the board.