NIGnets is a library for creating Neural Injective Geometry. Neural Injective Geometry is a method of geometry parameterization that has non-self-intersection as an in-built constraint. It is created through a specifically designed neural architecture. NIGnets is written in PyTorch and is therefore fully differentiable and can be used in shape optimization procedures.
These tutorials are meant to serve as documentation for the API and also give an indepth account of NIGnets as a concept by itself.