hosford42 / PPO

Reinforcement Learning

Resources

Requirements

Summaries

Demos

• Continous Problem
• Discrete Problem

Resources

• https://arxiv.org/pdf/1707.06347.pdf (PPO)
• https://arxiv.org/pdf/1506.02438.pdf (GAE)

Requirements

• python 3.x

Install

> pip3 install -r requirements.txt

Summaries

Start Tensorboard on the summaries directory, create one if there is none then run:

tensorboard --logdir=summaries

Demos

Solutions to two of the gyms from openAI

Continous

> python3 pendelum_demo.py -n -t (To train)
> python3 pendelum_demo.py -p    (To play trained model)

Discrete

> python3 cartpole_demo.py -n -t (To train)
> python3 cartpole_demo -p       (To play trained model)