MCTS Tetris

A bot that plays tetris using MCTS, with a CNN baked into the reward system. Originally forked from Tetris-AI

Warning: The source code was run in a Windows 10 Powershell environment. If run on Linux, the requirements might not work properly.

Make sure to initiate a venv and install the requirements.txt.

For further instructions, check the help menu with:

py run.py -h

Name		Name	Last commit message	Last commit date
graphs		graphs
logs/tetris-nn=[32, 32]-mem=20000-bs=512-e=1-20240403-160740		logs/tetris-nn=[32, 32]-mem=20000-bs=512-e=1-20240403-160740
mcts_best_sequence_cnn_b		mcts_best_sequence_cnn_b
mcts_best_sequence_cnn_b_small		mcts_best_sequence_cnn_b_small
mcts_best_sequence_cnn_f+b		mcts_best_sequence_cnn_f+b
mcts_best_sequence_cnn_f+b_fixed		mcts_best_sequence_cnn_f+b_fixed
mcts_best_sequence_cnn_f+b_fixed_small		mcts_best_sequence_cnn_f+b_fixed_small
mcts_best_sequence_rf		mcts_best_sequence_rf
models		models
reward_regressors		reward_regressors
.gitignore		.gitignore
README.md		README.md
dataset_manager.py		dataset_manager.py
dqn_agent.py		dqn_agent.py
dqn_reward_function.py		dqn_reward_function.py
hi.png		hi.png
hi2.png		hi2.png
montecarlo.py		montecarlo.py
numpyarrayencoder.py		numpyarrayencoder.py
random_forest.py		random_forest.py
report.txt		report.txt
requirements.txt		requirements.txt
results.svg		results.svg
run.py		run.py
test.py		test.py
tetris.py		tetris.py
train.py		train.py
tree_graph.png		tree_graph.png