Welcome

The goal of Jobman is to facilitate the process of launching many concurrent jobs, by automatically handling how parameters are passed to your programs and how results are stored for further analysis. While Jobman is application-agnostic, it is particularly well suited for machine learning when performing model selection (handling of hyperparameters, storing and analyzing results).

Jobman has two basic modes. It can be used directly from the command line (which will rely mostly on flat-files for data storage) or it can be combined with a PostgreSQL database for more intensive applications.

Documentation

Roughly in order of what you’ll want to check out:

• Installing Jobman – Installation instructions.
• Introduction – The basics of Jobman.
• Passing Parameters to Jobman – Learn how to use Jobman on the command line.
• Use Case: using SQL – Become familiar with SQL queries.
• Multilayer Perceptron with Jobman – A complete example, from start to end.
• Database Architecture – Quick tour of the database architecture underlying Jobman.
• Advanced – Advanced topics.
• Tips – Tips.

You can download the latest PDF documentation, rather than reading it online.