Error in DP-SGD

[apustozerova]

when trying to run Gradient Perturbation with logistic Regression or DP-SGD, when fitting the model with model.fit there pops up an error:

AssertionError: compute_gradients() on the differentially private optimizer was not called. Which means that the training is not differentially private. It happens for example in Keras training in TensorFlow 2.0+.