Skip to content

dletk/DeepNN2048

Repository files navigation

DeepNN2048

This is the code for my independent study project in Deep Neural Network, Fall 2017.

The project is written in Python 3 using the Keras library for building the neural network model. The model is trained and validated using an AI agent that I developed earlier. This AI agent applied the minimax algorithm with some heuristics to win the 2048 game. Using this version, I have generated 1000 games for the Neural Network to be trained on, and about 100 games to validate the network. The current model has the accuracy of 70%, making it achieve 1024 for 60% of the time.

In order to run the visuallization, please cd to /gameplay and run this command:

python3 GameManager_3.py

About

Deep Neural Network to solve 2048 in Python 3

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published