AlphaZero agent playing Losing Chess versus Human Player
- Clone or download repository
- Run command python3 main.py
- You will need to install several python libs in order to play the game: pygame, pytorch, python-chess
- Press ENTER to play with black pieces
- Press SPACE to play with white pieces
- Control number of MCTS sims, temp and cpuct param at the end of main.py file
- Currently AlphaZero agent is playing with 854th iteration of learning
- We added move evaluation based on stockfish engine (centipawn loss), and move prob chosen by AlphaZero to evaluate moves of AlphaZero agent
