chess-ai

Simple neural network implementation for chess in CLI

Requirements

pip3 install torch numpy python-chess

How to play:

python3 play.py

Use UCI format such as: g1f3, from square:(g1), to square:(f3)

Implementation:

• state.py returns a 8x8x5 bitvector, add an axis to the vector using .unsqueeze(0) then input it into the neural network in nn.py. Use the output for alpha beta in play.py.