Create a video of your reinforcement learning agent!

Start with step_1_train_model.ipynb to train the DQN network. Training will take days to weeks on a typical desktop. Or, skip to the step_2 notebook to learn how to generate a video.

Use step_2_evaluate_model.ipynb to generate a video with hidden layers outputs as 2d histograms. I included an example model 210612_1_SMB_DQN that can beat level 1-1.