Go to TogaWare.com Home Page. GNU/Linux Desktop Survival Guide
by Graham Williams
Duck Duck Go

Installing an MLHub Model/Package

20200221 After installing MLHub (Section 8.2) we are ready to install MLHub packages. The simplest option is to install curated packages. Such packages are reviewed by the MLHub AI team and the specific details required to install a package are obtained from a MLHub maintained index.

The list of curated packages is provided through the available command.

$ ml available
The repository 'https://mlhub.ai/' provides the following models:

animate      2.1.4  Tell a data narative through animations.              
audit        4.1.0  Classic financial audit predictive classification model.
azanomaly    3.1.4  Azure Anomaly Detection.                                
azcv         2.5.0  Azure Computer Vision.                                  
azface       2.0.8  Azure Face API demo.                                    

The simplest installation will then list one of the package names (from the left column of the above table). For example:

$ ml install rain

The pacakge itself is retrieved from a git repository, and from github by default. The file MLHUB.yaml will be obtained from that repository which contains the meta data required to install the package for MLHub. We can explictly identify the GitHub path, thus skipping the curated package list:

$ ml install gjwgit/rain

Specific branches and commits of a GitHub repository can also be identified. The example above accesses the master branch by default. The first example below accesses the latest dev branch. The second example accesses a specific commit.

$ ml install gjwgit/rain@dev
$ ml install gjwgit/rain@aace268

Specific MLHub yaml files within a github repository can also be identified:

$ ml install gjwgit/rain:doc/MLHUB.yaml
$ ml install https://github.com/gjwgit/rain/...

The default git repository is GitHub and so the above are the same as:

$ ml install github:gjwgit/rain

The open source GitLab repository is also supported:

$ ml install gitlab:gjwgit/rain@2fe89kh:doc/MLHUB.yaml
$ ml install https://https://gitlab.com/gjwgit/rain/...

As is BitBucket:

$ ml install bitbucket:gjwgit/rain
$ ml install https://bitbucket.org/gjwgit/rain/...

Copyright © 1995-2020 Togaware Pty Ltd
Support further development through the purchase of the PDF version of the book.
Graham Williams is the developer of open source software including rattle and wajig.
He is the author of Data Mining with Rattle and Essentials of Data Science.
This web site is hosted by Togaware, free and open source software since 1984.