PPO

Implementation of a multi-processed Proximal Policy Optimization (PPO) with some "implementation tricks" from the article What Matters In On-Policy Reinforcement Learning?

It achieves a score of around 300 (considered solved) on the BipedalWalker-v2 environment from OpenAI:

Update July 2021

Added the following features:

• Multi-processed workers for experience collection (resulting in 2-3x faster wall clock time).
• Model loading to allow warmstarting or continuing from a checkpoint.
• Proper seeding resulting in fully reproducible runs.
• Tensorboard logging.