Setup Keras-RL Deep Reinforcement Learning Environment on windows using pip

Source: Deep Learning on Medium


Go to the profile of Amir Khan

I assumed you already having Anaconda installed on your system, so lets start digging the further installations steps

Step 1 open anaconda command prompt, and on the terminal write the below command

pip install keras-rl
step 1

Step 2 , to run the keras-rl reinforcement learning examples, you’ll also have to install gym by OpenAI,using this link https://github.com/openai/gym#installation there are many installation steps, I will tell you which Installation step you follow, write the below command

pip install gym
step 2

along with gym you need to install one dependency h5py: simply run pip install h5py

step 2

basic Installation is been done for keras-rl reinforcement learning environment, for checking go to the python shell using python command and import gym

There is a special Installation for Atari Environment, which I will explain you how to install, For atari example you will also need Pillow

Install Pillow with the following command pip install Pillow

After that you can, use the following command to install Atari gym

pip install --no-index -f https://github.com/Kojoley/atari-py/releases atari_py

check atari

if you havE any dependencies error like i got missing cmake, make. Installed from the using for CMAKE link https://cmake.org/download/ use windows msi installer, for MAKE use http://gnuwin32.sourceforge.net/packages/make.htm , download make setup file and install, and and give the installed directory of make and cmake in the system environment PATH variable, to make it work on your cmd shell, like below

check cmake and make

https://stackoverflow.com/questions/42605769/openai-gym-atari-on-windows use the link for any help regarding error, it helps me!!

After Atari Environment Installed Successfully, Now the Fun Part!!!!!!!!

For running Atari Environment, open python shell and use below command

import gym
env = gym.make('SpaceInvaders-v0')
env.reset()
for _ in range(1000):
env.step(env.action_space.sample())
env.render('human')
env.close()
Atari Game

Have Fun with using other reinforcement learning environment and apply your learning on it!!!!, This is my first medium blog, And I am happy that I have made it more informative, THANKS:)