Fixes for GitHub Issue #979

[jamwar01]

• Root cause of issue was that per-channel clustering was being performed on a Conv2D layer with less weights per channel than the number of desired clusters per channel, leading to 1 clustered weight per channel in the resulting model
• Desired behaviour is that no clustering should occur in this scenario
• Behaviour also corrected to not perform clustering if too few weights in the cluster_per_channel=False scenario
• Warnings now given to user that a Conv2D passed to the clustering API had too few weights for the requested type of clustering and that no clustering was performed on it.
• ValueError given if layer passed to clustering API with no weights Change-Id: Idee3388e4905b5bf41f9598fa49ce200e9682bc6

[google-cla[bot]]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.