Added functionality to watch_agent.py to support evaluating different agents.

Added pull_config.py to grab configuration data from agent. Script is still being tested and tuned.

Added checkpoint agents to each dqn version. Added the notebook for dqn 2.5

Changed file structure for agents and updated readme

Updated tags in readme file

removed example video

record_video.py now functions, needs tuning

Modified the record_video.py file, still not functional.

Added the a2c notebook and the latest dqn run notebook

Updated the agent file and best model file for version_2

Added option flags to the watch_agent.py file.

Modified watch_agent.py to include ability to give an argument to adjust repeat action probability.

Added video recording, plotting, and viewing script templates.

Changed naming convention for agent directories and TensorBoard directories

Removed redundant ppo notebook

Added PPO trial run

Added run2v3 to notebooks. Updated run_2 model and best model. Added run2v3 TensorBoard and eval files to agents.

Added second Tensorboard file from run2 to its own file in order to split it out in the Tensorboard dashboard.

Updated run2v2 notebook

Updated run_2 agent and best model. Added evaluations and Tensorboard files to run_2.

Added final notebook files for first runs. Added Agents and evals from first runs.

Added dqn_pacmanv5_run2.ipynb

Fixed video callback freq in dqn_pacmanv5_run1.ipynb prior to training.

Updated dqn_pacmanv5_run1.ipynb to final version prior to training.

Added notebook structure for training first agent.

Updated the README

Added to the README

Added license

initial commit