Concatenating SVM with GRU model

Question

I am trying to concatenate GRU with SVM using numpy. I am currently using hidden layer calculated from GRU as input for SVM and calculate the true result, and it does work out pretty well. Is it possible for me using tensorflow 2.0 to do so?

Answer 1

If by 'hidden layer' you mean taking the internal state of the GRU, then it is supported by Tensorflow. What you need to do is setting return_state parameter to True when creating the GRU layer, as explained in the official documentation.