DQN plays game 2048 Author: Drake
Environment: game_2048.py
if you'd like to test(play) the game please try: play_human.py
(just use keyboard 'u','d','l','r' for up, down, left, right operation)
DQN: RL_brain.py
train the network: train_AI.py
test the pretrained model: play_ai.py
Better than inital model but Saddly, my trained model didn't play very well...(500 epochs, 32 batch/epoch)