raise error if tfp.layers' kernel_prior/posterior aren't specified with correct shapes

[dustinvtran]

A common user mistake is to specify the kernel_prior or kernel_posterior distribution without a matrix event shape. This makes the KL divergence computation incorrect as the KL must be over matrixvariate distributions. Let's catch this early on by checking during build().

Same applies to bias vector.

[LexusH]

Hello, I'm new to this project and would love to help with your recommended changes. Any guidance on this topic would be much appreciated, but I will go digging around in the files you brought up immediately to see if I can fix this!

[LexusH]

I have a small demo up to try and test the error you're mentioning, but I don't think I understand what output method to be using to see your error. I'd love to get your input on what i'm doing wrong here:
https://colab.research.google.com/drive/1qhKpW5FdD0bHvE6BOjoT5_hU7kkL6A_J