v0.2.0-publication

Version of the repository at time of publication, incorporating changes from the review-process.

What's changed.

-Interface have been changed to be more explicit in the use of finite-difference approximation, so as not to accidentally loose precision when using float64 in place of pytorch default float32.