Team members:
Om Khare (MIS 112003066)
Harshmohan Kulkarni (MIS 112003075)
The project aims to develop a reinforcement learning (RL) agent that can effectively play and potentially master the game 2048, a popular single-player puzzle game. The agent uses a Deep Q-Network (DQN) to evaluate and select the best moves based on the current state of the game board.
Use Version 3 to play 2048 with the improved model compared to prior versions.
Use v3_TrainAI.py to train the model to play the 2048 game. Use game_v2.py to play 2048 on the terminal.