gradient based SNN training in the complex plan

implement the diffeqs of SNN and see if one can avoid singularities by going to the complex plane and picking up principal values?