Running GibsonEnv on headless server

• How to run GibsonEnv in headless mode if I have X server running?

• Follow these steps:

  1. Install nvidia-docker2 dependencies following the starter guide. Install x11vnc with sudo apt-get install x11vnc.
  2. Have xserver running on your host machine, and run x11vnc on DISPLAY :0.
  3. docker run --runtime=nvidia -ti --rm -e DISPLAY -v /tmp/.X11-unix/X0:/tmp/.X11-unix/X0 -v <host path to dataset folder>:/root/mount/gibson/gibson/assets/dataset <gibson image name>
  4. Run gibson with python <gibson example or training> inside docker.
  5. Visit your host:5900 and you should be able to see the GUI.

• How to run GibsonEnv in headless mode if I don't have X server running?

• Follow these steps:

  1. Build your own docker image or pull from xf1280/gibson:0.3.1.
  2. Run with docker run --runtime=nvidia -ti --rm -p 5001:5001 -v <host path to dataset folder>:/root/mount/gibson/gibson/assets/dataset gibson Note that we don't need to set DISPLAY variable and pass /tmp/.X11-unix/X0 to container anymore, but we open 5001 port for visualization.
  3. In docker image, change the configuration file, set display_ui to false and mode to headless.
  4. open tmux, in one tab, run python <the gibson file you want to run; in another tab run python gibson/utils/web_ui.py 5552.
  5. See the rendering in host:5001