suragnair / alpha-zero-general

A clean implementation based on AlphaZero for any game in any framework + tutorial + Othello/Gobang/TicTacToe/Connect4 and more
MIT License
3.9k stars 1.04k forks source link

Actionspace 1 to many #305

Closed pieterjanpop closed 1 year ago

pieterjanpop commented 1 year ago

What is the reason that the action space is enlarged by one? Doesn't this lead to wrong models?

Schermafbeelding 2023-05-12 194708

zimpzon commented 1 year ago

It is for no move. Allowed in some games, not in others.

pieterjanpop commented 1 year ago

How do I handle this in a game that does not allow it?

zimpzon commented 1 year ago

You just leave out that last element. I can't help you more than that, I'm having some trouble implementing my own custom game. Don't expect it to be easy ;)