Add randomised benchmarking fidelity decay model

[amycxh]

This should probably subclass the exponential decay model with a change of base. I'm trying that now. But I know that this version is tested and works, at least, so it's a simple starting point.

@hartytp How attached are you to calculate_derived_params remaining a static method?

[amycxh]

I've made a PR https://github.com/OxIonics/ionics_fits/pull/55 to this branch to make the Benchmarking model inherit from Exponential and I think I prefer that version.

[amycxh]

Alright, I've decided to keep things simpler and use this version, with more natural fit parameters, but have rewritten to use y0 and y_inf like the Exponential model, rather than A and B.

[amycxh]

Thanks @hartytp - this was my first foray into ionics_fits and it was very easy to use 🙂

[hartytp]

Thanks!