Added API ood post-training ood classifiers

[albertogaspar]

This PR contains the methods introduced in Lee, Kimin, et al and Mukhoti, Jishnu, et al.

These are post-training methods that aim at fixing the overconfident behaviour of deep neural net on inputs that are far away from the training data.

Pull request type

Please check the type of change your PR introduces:

• Bugfix
• Feature
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no api changes)
• Build related changes
• Documentation content changes
• Other (please describe):

What is the new behavior?

• Introduced an Abstract Base Class for post-training methods that aims at fixing the overconfident behavior of deep neural net on inputs that are far away from the training data.
• Implemented Lee, Kimin, et al
• Implemented Mukhoti, Jishnu, et al.
• Added a new runnable example on 2 moons that make use of Lee, Kimin, et al.
• Merged the runnable example with the one for SNGP.

Other information