Gomoku_Reinforcement_learning

Gomoku AI designed for ESC190 competition

Methods:

Greedy search
Minimax
Reinforcement learning
- policy value network
- monte carlo search tree
- train via selfplay

Name		Name	Last commit message	Last commit date
Latest commit History 4 Commits
.gitignore		.gitignore
README.md		README.md
checkpoint2-policy-bn.h5		checkpoint2-policy-bn.h5
checkpoint2-selfplay-3000.h5		checkpoint2-selfplay-3000.h5
checkpoint2-selfplay-bn.h5		checkpoint2-selfplay-bn.h5
checkpoint2-selfplay-eval-100.h5		checkpoint2-selfplay-eval-100.h5
checkpoint2-selfplay-eval-200.h5		checkpoint2-selfplay-eval-200.h5
checkpoint2-selfplay-eval-300.h5		checkpoint2-selfplay-eval-300.h5
gomoku.py		gomoku.py
gomoku_MCTS_pvnet1.py		gomoku_MCTS_pvnet1.py
gomoku_MCTS_pvnet2.py		gomoku_MCTS_pvnet2.py
gomoku_Minimax.py		gomoku_Minimax.py
gomoku_Score.py		gomoku_Score.py
loss_data.pickle		loss_data.pickle
model.py		model.py
plot.py		plot.py
predict.py		predict.py
records.docx		records.docx
score_data.pickle		score_data.pickle
train_selfplay.py		train_selfplay.py

Provide feedback