There are a number of options for getting started with exploring MLHub. All of the curated models are currently tested against Ubuntu LTS (20.04 and 18.04) and should work out of the box on that platform. Ubuntu LTS is available everywhere and there are many options for accessing Ubuntu.

Options for Running Ubuntu

  1. Windows 10’s Windows Subsystem for Linux runs Ubuntu natively within a Windows 10 system. It’s a good choice.
  2. A Docker image is available for mlhub. The image is downloaded and run in a container under Linux, OSX, or Windows. See the Docker Hub for details.
  3. Install Ubuntu 18.04 LTS on your own computer as the base operating system. Many people run Ubuntu rather than Windows so you will be in good company.
  4. Run Ubuntu on Windows 10 using Hyper-V Manager, under the Quick Start choose Ubuntu 18.04 LTS to install a virtual server locally.
  5. On Macintosh OSX install Ubuntu under the Parallels virtual machine platform to install a virtual server locally.
  6. Fire up a Ubuntu 18.04 LTS server on Azure. You can get free credit (USD200) from Microsoft to trial Azure. Then in the portal add a new Ubuntu 18.04 LTS server, creating a new resource group for it, providing a name for the virtual machine, a user name and a password, choosing a location (e.g., westus2), opening network port 22 for ssh connectivity, and choosing a machine size to suit your needs. Once created go to the portal and provide a DNS name from the overview page’s, DNS link. Connect to the server using ssh from a terminal to the DNS name, set the server up as below, then connect from X2Go.

Once you connect to the Ubuntu system the following initialisations of the new server can be useful, and will take about 10 minutes, depending on network connectivity.

$ sudo apt-get install wajig
$ wajig update
$ wajig distupgrade
$ wajig install gnome-terminal ubuntu-mate-desktop x2goserver python3-pip

If you are connecting to a remote server then once the above is setup you can connect to the server using X2Go for a desktop experience, rather than via a terminal, using ssh -X.

We are now ready to start with mlhub:

$ pip3 install mlhub
$ ml                  # May need to log out and back in to find command.
$ ml configure
$ ml available

See the Quick Start Guide for details.