Running reinforcement learning lab on Windows

I am sharing my experience when attempting to run C3_W3_A1 assignment (the lunar lander) on my Windows PC.
Unfortunately, I discovered that this does not work, as the graphical renderer Display() from pyvirtualdisplay requires X.
I did various attempts to solve the situation, including:

  • cloning box2d and attempting to install it locally. It did not work because swig complaints about some directives.
  • putting the notebook in google colab. Unfortunately, it seems that xvfb is not available in this environment and cannot be installed.

Nothing worked, so I suppose my last resort could be a VM or a disk partition with Linux - or maybe run the lander “headless” (no graphics).
If someone has a better solution, I would be grateful to get hints!

It is difficult because many of the packages used in that lab are no longer supported.

You might find other threads on this topic by using the forum search tool for the word “lander”.

1 Like

Thanks for your answer.
Yes, I found in the forum other people with the same problem - and no solution it seems.
Too bad, but after all the lab is about DQN… I’ll go headless or just pass :grinning: