Reinforcement Learning Project
This project is an attempt to learn PPO (Proximal Policy Optimzation) with a complex robotic arm environment.
Files of note:
train_arm.py
- with some modification, it will kick off training a configurable environment of the robotic armplayback.py
- with some modification, it will load a given agent's model file and execute it on a given environment for several episodes, then save the results to a.gif
file.agents
folder has all trained agents for given environment configurations and experiments