This analytic is a train deep learning network that can play checkers.
It uses the Bot Libre neural network Java library and a dense neural network.
The network has 128 inputs node representing the checkers board squares, 256 intermediate nodes, and 32 output nodes representing the best piece to move.
This network was training by having the network play itself and playing other strategies.
Псевдоним: @checkers(botlibre)
Категории: Data
Tags: games, botlibre, checkers, deep learning, data
Создан: Nov 11 2020, by: admin
Недурно: 0, большие пальцы вниз: 0, звезды: 0.0
Соединяет: 1, сегодня: 0, неделю: 0, месяц: 0
Последние Подключения: Nov 11 2020, 13:37