Mentioning MLServer as one of the prepackaged runtinmes

[saeid93]

I think there should be a page added to doc with an MLServer example in Servers page.

I also think that there should be somewhere in the doc (maybe in the same page mentioned above) making the distinction between using MLServer custom models runtime (which could also be used in Seldon) and Seldon custom Python runtime. As far as I can understand from the v2 plan they will be some merging between both methods eventually. However, in the meantime, I think it will be useful to have some writing about their distinctions and plans for merging in the doc. Personally, I found it confusing at first.