Make sure that when you run
docker ps on the command line you get
output similar to the following (if not, then either Docker is not
installed, is not running, or your user does not have priveleges to
access the Docker group). OpenAI Universe will not run if this
step is not met.
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
Ensure you are using the virtualenv you created for this project, you can go into it by running the following (substituting for where you located your virtualenv.
Go into an interactive python shell by typing the following on the command line:
Copy and paste the following lines of python code:
# THIS CODE IS TAKEN FROM THE OPENAI GITHUB PAGE # https://github.com/openai/universe#run-your-first-agent import gym import universe # register the universe environments env = gym.make('flashgames.DuskDrive-v0') env.configure(remotes=1) # automatically creates a local docker container observation_n = env.reset()
Note: You may need to hit enter one or two extra times after pasting this into the interactive shell to make sure the last line of code gets executed.
When you first run this, it will take some time to download all the files for that game, and you will see progress bars as it does this.
After it has done this, it will continually spit out a large amount of
server log printouts about
clients. At this stage it should automatically open
up an internet browsers. Within your browser window, you will see
screen capture of a chrome browser, playing a flash game, like in
the following image.
If it does not open up a browser automatically (it didn't for me), then you will need to manually enter the following URL in your internet browser:
While it is still spitting out text on the printout, paste the following lines of code in one go (again you may need to press ENTER one or two extra times to get the final line to go through).
# THIS CODE IS TAKEN FROM THE OPENAI GITHUB PAGE # https://github.com/openai/universe#run-your-first-agent while True: action_n = [[('KeyEvent', 'ArrowUp', True)] for ob in observation_n] # your agent here observation_n, reward_n, done_n, info = env.step(action_n) env.render()
You should hopefully start seeing, seeing some printouts on the
command line about
rewards. Also, you should see the car
in the window actually driving around by itself.