Update to newer GPflow and TensorFlow.

[jesnie]

Update to make GPflux compatible with newer versions of GPflow, that require an additional X parameter in the likelihoods.
In this PR I just pass a dummy None value as X. Alternatively we could:

1. I don't know Keras and GPflux well, but maybe we can find a "real" value of X to use?
2. Do we want to attempt to write code that's compatible with earlier version of GPflow as well? I suppose we could add an if somewhere?

[SebastianPopescu]

all seems good from an initial view.

[SebastianPopescu]

I think passing None is still not best practice as if someone wants to actually build a heteroskedastic model in GPflux they are gonna have a harder time to modify stuff in the lack of a blueprint, but it does the job for now.