Provided dockerfiles (all images use Ubuntu 16.04):
- python3 - Python3, ViZDoom with dependencies and X11 support. By default runs Python basic example.
- tensorflow - Python3, ViZDoom with dependencies, Tensorflow and X11 support. By default runs tensorflow learning example.
- torch - Torch, ViZDoom with dependencies and X11 support. By default run Lua basic example.
# Uses Dockerfile in python3 directory to create image tagged vizdoom
docker build python3 -t vizdoom
# Uses Dockerfile in tensorflow directory to create image tagged vizdoom_tf
docker build tensorflow -t vizdoom_tf
# '--net=host -e DISPLAY=${DISPLAY} --privileged' required for GUI
# Run basic example in container named 'basic'
docker run -ti --net=host -e DISPLAY=${DISPLAY} --privileged --rm --name basic vizdoom
# Run tensorflow learning example in container named 'vizdoom_tf'
nvidia-docker run -ti --privileged --net=host -e DISPLAY=${DISPLAY} --rm --name vizdoom_tf vizdoom_tf
# Run bash in interactive mode
docker run -ti --privileged --net=host -e DISPLAY=${DISPLAY} --rm --name basic vizdoom bash
- Ubuntu 17.04, Anaconda3-4.4.0 by @guiambros