kayon/mlhub repository overview

Please star the image if you find it useful. Thanks.

This is a fully functional standalone MLHub (https://mlhub.ai⁠) on a Ubuntu server ready to run on Linux, Mac OS X, and Windows. Many common ML model dependencies are installed and others are installed on demand.

To start running the image see the commands below, and once you start running the image a bash shell is started and there you can type the following commands:

$ ml available
$ ml install objects
$ ml configure objects
$ ml demo objects
...

See https://mlhub.ai⁠ for details.

To start running a docker image in a container on your operating system, do one of the following, as appropriate.

⁠GNU/Linux

$ xhost +
$ docker run -it --rm --network=host --env="DISPLAY" -v ~/.Xauthority:/root/.Xauthority:rw mlhub

The xhost command avoids the error you might get as an XWindows permission error. The xhost command is the "nuclear" option to give permissions, before running the docker command.

⁠Mac OS X

There is some helpful information at (https://fredrikaverpil.github.io/2016/07/31/docker-for-mac-and-gui-applications/⁠).

1. Install Docker and XQuartz⁠.
2. Obtain the relevant Docker image (optional as docker run will automatically download the image).

$ docker pull kayon/mlhub

1. Open XQuartz and then in preferences under the Security tab tick "Allow connections from network clients"

$ open -a XQuartz

1. Find out your IP address and allow xhost permissions

ip=$(ifconfig en0 | grep inet | awk '$1=="inet" {print $2}')
xhost + $ip

1. Run the Docker image in a container

docker run -it --rm -e DISPLAY=$ip:0.0  kayon/mlhub

⁠MS/Windows

1. Install Docker for Windows⁠ and then Xming X server for Windows⁠.

2. Obtain the relevant Docker image (optional as docker run will automatically download the image).

$ docker pull kayon/mlhub

1. Use XLaunch to start an X server in Windows, with "No access control" box checked in the last step of launching.

2. Run the Docker image within a container providing the IP address of the host server (not the DockerNAT) as reported by the 'ipconfig' command under Windows.

$ docker run -it --rm -e DISPLAY=<ip>:0.0 kayon/mlhub