Allow learning_rate=0

[drasmuss]

Motivation and context:
Often when working with a learning model it is helpful to quickly disable learning by setting the learning rate to 0. Currently we don't allow this in Nengo, which isn't the end of the world but just makes things a little awkward. I think there is pretty minimal downside to allowing it; it seems unlikely that users will accidentally set the learning rate to zero when they do actually want learning.

How long should this take to review?

• Quick (less than 40 lines changed or changes are straightforward)

Types of changes:

• New feature (non-breaking change which adds functionality)

Checklist:

• I have read the CONTRIBUTING.rst document.
• [n/a] I have updated the documentation accordingly.
• [n/a] I have included a changelog entry.
• [n/a] I have added tests to cover my changes.
• [n/a] All new and existing tests passed.

[studywolf]

this has come up for me a bunch of times would be very handy

[jgosmann]

LGTM. Simulations are probably faster when deleting the learning rule, but that is definitely more awkward for quickly disabling learning rules.

[tbekolay]

LGTM, I can't think of a non-stubborn reason not to allow this. Especially when thinking about a situation where running an experiment with a bunch of different learning rates, including 0 to get a baseline error rate measurement.

I'll add a changelog entry in the merge, as this does affect end users.