Minor suggestions for make_perturbed_fun

[carlosgmartin]

I have some minor suggestions for make_perturbed_fun:

• Rename sigma to scale, which is more generic and self-explanatory. Note that Berthet et al. sometimes uses $\varepsilon$ for this, instead of $\sigma$.
• Change the default value for num_samples from 1000 (a large arbitrary value) to 1. This is a less arbitrary value, and more common in stochastic optimization (where the objective function is itself stochastic).

If one or both of these are desirable, I can create a PR. Otherwise, this issue can be closed.

[vroulet]

I like both suggestions. @q-berthet ok with you?

[YashSachdeva]

Hi @carlosgmartin @vroulet @q-berthet, I'd love to contribute and can implement both suggested changes if you're okay with it. Let me know if I should go ahead and open a PR! 🚀

[haroon0x]

I can work on this.