[DOC] Tutorial could include a classification example

[bmcfee]

One of the common questions that I get is "how should i build an audio classifier?"

My canned response is "openl3 + sklearn RandomForest". The tutorial in the docs is great for the first half, but if it was expanded to show how to combine it with a classifier on some example dataset (eg urbansed or something), it would be much easier for novices to pick up and run with.

[justinsalamon]

Great suggestion. Feel like making a PR? Otherwise we'll try to get to this, cycles permitting.

[bmcfee]

I'm too slammed to commit to a PR right now. It's not a pressing issue; I just wanted to flag it to be done whenever's convenient.

[pohpas]

@bmcfee do you have, by chance, a minimal example on how you added a classifier?

[bmcfee]

@pohpas sorry, I don't - this is yet another github issue from the Before Times™ that I haven't circled back to yet. Probably the simplest thing to do is adapt the openmic baseline notebook to use openl3 instead of vggish: https://github.com/cosmir/openmic-2018/blob/master/examples/modeling-baseline.ipynb