|
This program is a console mode program. When executing the program, you will see the window as below:
If you choose y, then the program will try to load TDL1A.sav and TDL2A.sav and use the data in the file to initialize the neural networks, otherwise the program will use some default value to initialize the neural networks. After choosing not to load data from files, you will encounter the window below:
If you choose y, then you can play with the neural network. Because the neural network is not very smart, the neural network will always the player that can put his piece first. If you choose n, then the program will train two neural networks by playing with each other. The number of episodes will be determined by you.
The training window will go on as below:
After input the number of episode you want to train, the program will go on training and then output TDL1A.sav, TDL2A.sav and Table.sav to show the training result. After that, the program terminates.
The playing window will go on as below:
The neural network always plays first, and then its your turn to put your piece now. The position you input will like this:a b, where a and b are integers from 0 to 8. If the number you input is wrong, you will be asked to input the position again.
Most of the time, you will win the neural network like this...
If you press n, the program will then terminate, otherwise, the new episode will begin, and still, the neural network plays first.
Tutorials
|
|