I'M Trying To Run A Reinforcement Learning Task With Trains-Agent, And The Environment I'M Running Needs An X Server To Get The Environment State. I Can Get Around This Using

Answered

I'm trying to run a reinforcement learning task with trains-agent, and the environment I'm running needs an X server to get the environment state. I can get around this using xvfb-run -s "-screen 0 1400x900x24" python task.py when running the task from command line. Is there a way I can get trains-agent to do the same?

  				
Posted 
	4 years ago

					More
				  		
  Report
		
					PunySquid88
				
					0
					 × 1

Votes Newest

Answers 3

Hi PunySquid88 , you can do that by running trains-agent using the same xvfb-run command, for example xvfb-run -s "-screen 0 1400x900x24" trains-agent daemon ...
When this trains-agent instance will start running an experiment in a python virtualenv, all environment variable will be visible to your experiment code...

  				
Posted 
	4 years ago

					More
				  		
  Report
		
					SuccessfulKoala55
				
					0
					 × 1

awesome thanks!

  				
Posted 
	4 years ago

					More
				  		
  Report
		
					PunySquid88
				
					0
					 × 1

PunySquid88 you can also run xvfb directly from python https://pypi.org/project/xvfbwrapper/ . We have been using it this way.

  				
Posted 
	4 years ago

					More
				  		
  Report
		
					NutritiousBear41
				
					0
					 × 1

Write your answer

1K Views

3 Answers

4 years ago

one year ago