1. pip install -r requirements.txt to install dependency.
2. Install pytorch.
3. pip install . to install this package.

• The ability to specify arguments in a file, and then can be optionally overridden by CLI.
• The ability for fast prototyping by declaring arguments locally.
• The ability for access the params everywhere (i.e., a global entry).

Arguments should be declared where they belong. For instance, arguments related to training should be declared in a Trainer class.

A library built upon arglib and trainlib which helps develop ml models.

Just a collection of useful functions and classes to deal with training.