Does the training of majority voting in scikit-learn will re-train the classifiers?

Question

I am curious whether the training of majority voting in scikit-learn will re-train the classifiers?

For example:

model_perceptron = CalibratedClassifie