- Update to new stdlib organization.
- All properties in a single file, with formulation closer to the original.

Avoid splitting when not necessary as it may incur in performance hit. Indeed,
splitting means more (tinier) goals to verify, hence more call to provers.

Instead of Classifier.

Also update the codebase accordingly.

Modify the printer and transformations for Marabou to accept working on reals.

François Bobot authored 2 years ago and Michele Alberti committed 2 years ago

Make it into a new strategy module that should prepare the task before sending
it to the provers.

- Add support for neural network as array in language.ml
- Translate csv record as vnnlib constraints in dataset.ml
- Add simple dataset normalization functionalities in dataset.ml
- Add vnnlib printer (as smtlib printer with direct real support)
- Add vnnlib driver for PyRAT
- Rework onnx and nnet parsers to not fail on non-mandatory information

The driver seems to do the job already.

Include building a proper Why3 float literal.

François Bobot authored 2 years ago and Michele Alberti committed 2 years ago

The axiom a \/ b gives two different tasks.

           - more robust
           - fix a perhaps why3 change