JJukE / ReinforcementLearning

Reinforcement Learning Course

ReinforcementLearning

Reinforcement Learning Course

Docker - RL & openAI Gym

Description:

A docker environment for RL in python 3 including the OpenAI Gym toolkit INcludes:

1. Basics: NumPy, Pandas, Scipy, Jupyter, Matplotlib
2. Deep Learning: Tensorflow, Keras
3. Reinforcement Learning: Keras-RL, baselines, TensorForce
4. Environments: AI Gym
5. Others: ipywidgets, h5py

Docker Hub

You can directly pull the built image from Docker Hub by running

docker pull jaimeps/rl-gym

Rendering on Jupyter notebook

The virtual frame buffer allows the video from the gym environments to be rendered on jupyter notebooks. Simple example with Breakout:

import gym
from IPython import display
import matplotlib.pyplot as plt
%matplotlib inline

env = gym.make('Breakout-v0')
env.reset()
for _ in range(1000):
  plt.imshow(env.render(mode='rgb_array'))
  display.clear_output(wait=True)
  display.display(plt.gcf())
  env.step(env.action_space.sample())

Reference

github : https://github.com/jaimeps/docker-rl-gym