kongjiellx / AlphaZero-Renju

AlphaZero-Renju

References：

1. A blog explain how mcts works
2. AlphaZero connect four, there are someting interesting ideas, like use Q to instead of Z to be training target.
3. Some notices about mcts
4. tensorflow AGZ implement(NOT an official version of AlphaGo)
5. A renju program follow AlphaZero. Here is a chinese doc about it. I think this implemention has some bugs. For example the author compute final z by mistake, but when do backup he adds a negative sign, so maybe the result is ok.

Requirements

1. This project build with bazel==2.0.0
2. python requirements

Results

• Training about 24 hours, each move with 100 mcts simulations:

• Training about 48 hours, each move with 400 mcts simulations:

test