Implementation of Logic Tensor Network with pytorch.

Trying to follow 10.1016/j.artint.2021.103649

examples based on https://github.com/logictensornetworks/logictensornetworks/tree/master/examples

• parsing axioms (str with real logic -> torch model - like object)
• domain control (variables shapes checking, type casting)
• constants
• choice of fuzzy logic (T-norm, conorm)
• variables values cross product
• Diag operator
• aggregation operators
• Guarded quantifiers
• allowing not torch.nn.Module to be predicates of functions
• calculating SatAgg