all essentials for OpenAI gym to work with NVIDIA GPU, using tensorflow/tensorflow as base image.
2.1K
This docker image adds all the required packages on tensorflow/tensorflow:latest-gpu-py3-jupyter image to make OpenaAI gym, GUI work with NVIDIA GPU inside the container.
To get started, make sure to go through this tutorial to install nvidia-docker2 and tensorflow/tensorflow:latest-gpu-py3-jupyter.
To make openGL, gedit and other GUI applications work
Enter the following command in the host PC
sudo xhost +local:docker
it isn't very secure but it is easy. If you want it to be more secure instead please, checkout this article
Don't forget to enter following after use
sudo xhost -local:docker
use the tags listed in X-server sub-section of usage section with the docker run command
use the tags listed in GPU sub-section of usage section with the docker run command
to make jupyter work on your host's browser use tags listed in jupyter sub-section of Usage Section.
In order to run this container you will need nvidia-dokcer2 and tensorflow/tensorflow installed on your system. Please, go through this article to install.
To check if it installed correctly
docker run --rm tf-gym:latest
To run interactive shell
docker run -it --rm tf-gym:latest bash
--volume="$HOME/.Xauthority:/home/tf/.Xauthority:rw" --env="DISPLAY"
-e NVIDIA_DRIVER_CAPABILITIES=all --gpus all
--net=host
Content type
Image
Digest
Size
3.6 GB
Last updated
over 6 years ago
docker pull k4droid3/tf-gym