mismatch between the documentation and the behaviour of `make_private_with_epsilon`

[omarfoq]

PrivacyEngine.make_private_with_epsilon does not behave as explained by the documentation. In particular, poisson_sampling can not passed as an argument to this method. I think that the error is due to calling get_noise_multiplier (in accountants/utils.py) with **kwargs, but RDPAccountant.get_epsilon (in rdp.py) and GaussianAccountant.get_epsilon (in gdp.py) only accepts a fixed number of arguments. Also, I think that the signature of GaussianAccountant.get_epsilon should be get_epsilon(self, delta: float, poisson_sampling: bool = True) -> float

[FrancescoPinto]

Actually the msimatch also includes the impossibility of choosing the clipping strategy with make_private_with_epsilon (as the parameter is not passed correctly to make_private)