Can weightwatcher still be useful for model with RNN layers?

[bkocis]

From the description it is clear that weightwatcher support only Dense / Linear / Fully Connected (and Conv1D) and Conv2D.

In the case of models that have any number of non supported layer (for example an LSTM) in a model - can the output of the ww analysis still be useful?
Can we make a judgement about such a model if WeigthWatcher analyses only the supported layers in a model?